Because Apple has not officially announced anything yet, all of these details should be treated as expectations rather than confirmed specifications. Still, multiple reports point in the same direction: the next MacBook Neo will be an iterative but important update that improves performance, raises the baseline memory, and helps bring Apple’s newer artificial-intelligence features to a more affordable machine.

MacBook Neo 2 features

Based on reported details and Apple’s typical product strategy, the MacBook Neo 2 could arrive with improvements spanning chip, memory, software, and design. Here is an in-depth look at everything rumored about the second-generation MacBook Neo.

A19 Pro chip and performance

The current MacBook Neo runs on the A18 Pro chip, the same processor used in the iPhone 16 Pro generation. For the next-generation model, Apple is reportedly working on a version powered by the A19 Pro chip. That will put the MacBook Neo in the same silicon family as the next iPhone Pro models and a new iPhone Air.

The performance jump is expected to be noticeable. The A19 Pro is described as offering roughly 10 to 15 percent faster single-core and multi-core CPU performance than the A18 Pro. While that kind of year-over-year CPU gain is not dramatic, it is still a solid improvement for an entry-level laptop. The graphics story is more significant. Benchmarks reportedly show GPU performance improvements of up to 40 percent, helped by a new architecture in which every GPU core includes neural accelerators. That design enables better parallel processing and should make the chip more efficient when handling graphics and AI tasks at the same time.

Apple will likely use a binned version of the A19 Pro in the MacBook Neo. Chip binning is a standard manufacturing practice in which processors that do not fully meet the highest specification are configured with slightly fewer active cores. In this case, the version of the A19 Pro used in the iPhone is expected to have a 6-core GPU, while the MacBook Neo’s version could be limited to a 5-core GPU. For practical use, that difference is unlikely to matter for the MacBook Neo’s target audience. The machine is not designed for heavy video editing or hardcore 3D graphics work. Instead, it is meant for everyday tasks such as browsing, writing, schoolwork, and light content creation, and the 5-core GPU should still deliver a substantial upgrade over the current model.

The chip choice also matters for software features. Apple has increasingly made its newest intelligence features depend on the processing power inside the device. Moving from the A18 Pro to the A19 Pro should allow the MacBook Neo 2 to support more advanced on-device AI models. This is part of Apple’s broader strategy of controlling the hardware and software stack so that new silicon can unlock experiences that would not work smoothly on older specifications.

Memory upgrade to 12GB

One of the few compromises in the original MacBook Neo is its 8GB of RAM. The rest of Apple’s Mac lineup now starts with 16GB of memory, which has made the 8GB configuration on the MacBook Neo stand out as a budget trade-off. Even so, the next MacBook Neo is not expected to make a full jump to 16GB. Instead, it is expected to offer 12GB of RAM.

A move from 8GB to 12GB may sound modest, but it should make a real difference in daily use. More memory allows users to keep more browser tabs open, switch between larger documents, run multiple messaging and productivity apps, and handle increasingly complex web pages without as much slowdown. For an entry-level laptop, 12GB is a balanced configuration that gives buyers meaningful headroom without pushing the price too high.

Memory also matters for on-device AI. Apple has repeatedly said that powerful local intelligence requires sufficient memory because models need to stay resident and responsive without constantly reloading. By upgrading the MacBook Neo to 12GB, Apple can offer a more capable set of AI features without increasing the device’s cost as much as a 16GB configuration would. For users who found 8GB limiting, this change is expected to be one of the most welcome improvements in the MacBook Neo 2.

New on-device AI features

The combination of the A19 Pro chip and more memory should allow the MacBook Neo 2 to support Apple’s more powerful on-device AI model. This newer model was described as enabling more accurate dictation and new options for customizing Siri’s voice. These features may sound small, but they point to a larger shift in how Apple is handling intelligence.

Dictation is already important for people who compose long emails, take notes by voice, or prefer to speak messages instead of typing them. A more accurate on-device dictation system would make the MacBook Neo 2 more useful in professional and educational settings. Siri voice customization, meanwhile, gives users more control over how the assistant sounds, making it feel less generic and more personal. Both of these features rely heavily on the neural engine and memory capacity.

The current MacBook Neo, with its A18 Pro chip and 8GB of RAM, reportedly cannot support Apple’s more powerful on-device model. This means the second-generation model is an important upgrade for people who want the cheapest powerful Mac but do not want to be left out of Apple’s newest AI features.

Colors and design

The original MacBook Neo is available in a set of bright colors, including citrus, blush, indigo, and silver. That palette distinguishes it from Apple’s more professional notebooks, which typically use muted tones such as space gray, silver, midnight, and starlight. The colorful approach has helped Apple market the MacBook Neo to students and first-time Mac buyers.

According to reports, Apple plans to update the MacBook Neo with new colors on a periodic basis. Rather than giving the laptop a new industrial design every generation, Apple may keep the frame largely similar and introduce fresh color choices to keep the product feeling current. This approach has worked well in other Apple product lines and makes sense for an affordable laptop that is expected to have a longer lifecycle. There are no details yet about which colors will be offered with the MacBook Neo 2, but it is safe to expect at least a change from the current lineup.

No major external redesign has been reported for the MacBook Neo 2. The laptop’s compact aluminum body, display size, and overall shape are likely to carry over, with the most significant changes coming from the internal hardware, memory, and software features.

MacBook Neo 2 release date

The first MacBook Neo was released in March. The second-generation model is not expected to follow the usual 12-month cadence. Instead, reports indicate that the A19 Pro version is being prepared for release sometime in 2027. That means Apple will likely continue selling the current model through the rest of the calendar year and into the following year before launching the successor.

A 2027 release window gives Apple time to refine the chip, memory configuration, and color options. It also aligns the MacBook Neo 2 with the broader A19 Pro product cycle across Apple’s iPhone and Mac lines. By the time the MacBook Neo 2 arrives, software developers will have already had time to adapt their apps to the A19 Pro architecture, and Apple’s operating systems will be better optimized for the device’s improved AI performance.

Apple has not confirmed a release date, final specifications, or pricing for the MacBook Neo 2. Until the company makes an official announcement, all reported details should be treated as early information from industry sources. Even so, the early picture is clear: the MacBook Neo 2 does not need a dramatic redesign to be a worthwhile purchase. A faster chip, increased memory, and the new AI features made possible by those changes would be enough to make it the most compelling entry-level Mac on the market.

Source: 9to5Mac News

This year’s September keynote carries extra weight. It will be John Ternus’s first major product launch since taking over as Apple’s CEO. Ternus, who previously led hardware engineering, has said that the event will showcase a major step forward for Apple’s lineup. The company is under pressure to maintain its momentum in an increasingly competitive market, and the products expected on stage cover Apple’s most important categories: the iPhone, the Apple Watch, and the AirPods. Apple’s September event is not just about hardware; it is also expected to introduce new versions of iOS, iPadOS, and watchOS. The company is likely to release release candidates to developers immediately after the event, with public versions following a week or so later.

iPhone

Apple is expected to announce three new phones. The iPhone 18 Pro and iPhone 18 Pro Max will be the flagship models, powered by the next-generation A20 Pro chip. Rumors point to a new camera system with improved low-light performance, larger sensors, and an advanced telephoto lens. The Dynamic Island is also shrinking, freeing up extra screen space without losing Face ID or camera capabilities. The Pro models are likely to arrive in new colors, possibly with a more scratch-resistant finish.

• iPhone 18 Pro: A20 Pro chip, major camera upgrades, new colors, smaller Dynamic Island.
• iPhone 18 Pro Max: Larger battery and better efficiency for the best battery life ever in an iPhone.
• iPhone Ultra: Apple’s first foldable iPhone, with a book-style display and premium materials.

The iPhone Ultra is the most intriguing addition. A foldable iPhone has been rumored for years, and its arrival represents a significant shift in Apple’s smartphone strategy. The device is expected to feature an exterior display and a large, crease-resistant inner screen. It could be sold alongside the Pro models rather than as a replacement, giving Apple a third tier in its lineup. Pricing is expected to start around $1,999, making it the most expensive iPhone ever. The foldable form factor could open up new possibilities for multitasking and media consumption, especially if Apple’s software makes efficient use of the larger canvas.

Apple Watch

Apple Watch is always a fixture of the September event. This year, two new models are expected: Apple Watch Ultra 4 and Apple Watch Series 12. Both should bring more powerful chips and new health features. The Ultra 4 is built for athletes and outdoor enthusiasts, so its upgrades may include a brighter display and a new health sensor for altitude or water temperature. The Series 12, meanwhile, is aimed at a wider audience. Reports suggest an always-on heart monitor that can track heart rhythm every five seconds, even at night. A ceramic shell option is also reportedly back, after being absent from many recent generations, giving Watch buyers a luxury alternative to stainless steel or aluminum.

• Apple Watch Ultra 4: New health sensors, design updates, and a faster processor.
• Apple Watch Series 12: Upgraded S-class chip, always-on heart monitor, ceramic option.

Apple’s wearables strategy has focused on health and safety features. The company has been obtaining regulatory approvals for new capabilities, and the Ultra 4 could include a sleep apnea alert similar to the feature added in recent generations. Another potential addition is a body temperature sensor for women’s health. These are the kinds of updates that encourage existing users to upgrade. With watchOS 12 likely to be released alongside the new hardware, Apple is expected to introduce additional fitness metrics and deeper integration with the Health app.

AirPods

AirPods 5 will be the sixth major new product to launch at the event. According to leaks, Apple will offer two versions: a standard AirPods 5 and a higher-end model with Active Noise Cancellation. This mirrors the current AirPods 4 lineup, which Apple introduced with mid-cycle updates. The new models may include an updated chip with improved Bluetooth performance, enriched bass response, and a longer battery lifespan. The updated AirPods will arrive alongside iOS 26 for iPhones and Apple Watches, ensuring that the new audio features work seamlessly with the latest devices. Apple’s focus will be on easing pairing, improving call quality, and reducing background noise during phone conversations. Rumors of a camera-equipped AirPods product are premature; informed reporting says that product is destined for 2027, not 2026.

• AirPods 5: Updated base model and

  
  Source: 9to5Mac News

FAA confirms overrun during landing

According to the Federal Aviation Administration, the accident happened shortly after touchdown on the afternoon of Sunday, Sept. 6, 2026. The agency said in a statement:

21 Air Flight 7598 overran the runway after landing at Miami International Airport around 2 p.m. local time on Sunday, Sept. 6. The Boeing 767-300 cargo aircraft departed from Luis Muñoz Marín International Airport in San Juan, Puerto Rico.

The FAA's language confirms that the aircraft had already touched down and was trying to bring the landing to a safe stop when it left the runway. Runway overruns can occur when an airplane lands too fast, too deep into the runway, or when braking is compromised by wet or contaminated pavement or mechanical problems. It is too early to say what role, if any, those factors played.

A large emergency response

Miami-Dade Fire Rescue deployed more than 60 units to the site, and the airport stopped all departures and arrivals while emergency crews worked. Videos and photographs from near the crash showed the Amazon logo visible on the aircraft's tail, with black smoke pouring off the airframe and visible from the road. The plane came to rest beyond the airport boundary, a scenario that made the scene even more hazardous for responding firefighters because the accident site included the vehicles and the surrounding roadway.

Firefighters extinguished the flames and attempted to account for every person who had been on the aircraft and in the vehicles that were hit. At publication time, no list of crew members or confirmed fatalities had been released. Commercial flights from Miami, a major travel hub, were temporarily grounded, causing cascading delays across the terminal. Airport and county authorities were telling travelers to monitor their airline for revised schedules.

Amazon and 21 Air respond

Amazon quickly issued a statement through spokesperson Kelly Nantel, confirming that the aircraft belonged to Amazon Air and was operated by 21 Air. Nantel said the company was cooperating with authorities and focusing on the people involved. The statement read:

“We can confirm that

Source: The Verge News

A new kind of boox device

Boox is a well-known name in the e-reader and digital note-taking space, often competing directly with Amazon's Kindle lineup and Kobo devices. Most of Boox’s existing products run on Android, which provides flexibility to install third-party apps like Kindle, Kobo, and Pocket. The Picco, however, breaks from that tradition. It will not run Android. Instead, it uses a stripped-down, Linux-based operating system. That decision positions the Picco as a dedicated reading device closer to the original Kindle philosophy, rather than a multipurpose Android tablet with an e-paper screen.

Boox has not yet explained exactly which Linux distribution or kernel the Picco will use, nor whether the firmware will be open to the community. The company includes a few small utilities like a timer, a to-do list, and a reading tracking app. But for advanced users who enjoy tinkering with their gadgets, the Picco may feel restrictive compared to some alternatives.

The Xiteink X4 comparison

The Picco is commonly described as Boox’s take on the Xiteink X4, a similar ultra-small e-reader that has attracted attention for its portability and hackability. Both devices share the same 3.97-inch e-paper display size, which makes them noticeably smaller than a standard Kindle or Kobo. This size places them in a niche category often described as pocket readers or keychain e-readers. They are designed for quick reading sessions on a commute or while waiting, rather than long immersive reading experiences.

One of the most obvious differences between the Picco and the Xiteink X4 is the absence of magnets on the back of the Picco. Xiteink devices include magnets that allow the e-reader to attach to the back of a smartphone or other metal surface. This magnetic attachment has been one of the more distinctive features of the X4, since it lets users carry the reader with their phone or stick it to a locker or refrigerator. Boox has decided not to include that feature, which may disappoint fans of the original concept.

It is unclear why Boox omitted the magnets. It could be a cost-cutting measure or a design choice aimed at keeping the device lighter and simpler. Alternatively, Boox may have had concerns about interference with other electronics or durability issues related to repeated magnetic attachment. Without an official explanation from Boox, consumers are left to speculate.

Display and controls

The Picco’s 3.97-inch display is a touchscreen, which gives it an advantage over some earlier models in this category. A touchscreen allows for easier navigation through menus, text selection, and dictionary lookups. It also makes adjusting settings and browsing the small library more straightforward, especially when compared to a purely button-driven interface.

Physical page-turn buttons are present on the Picco, as they are on most e-readers. What sets the Picco apart is that these buttons are customizable. Users can assign different actions for short and long presses, potentially allowing one button to turn pages forward, the other to turn back, and a long press to jump to a chapter or open a menu. This level of control is unusual in such a compact device and should please readers who care about ergonomics and one-handed use.

Storage and file transfer

One of the most striking details about the Picco is that it appears to have no onboard storage whatsoever. Instead, the device relies entirely on a microSD card to hold books. This is a significant departure from most modern e-readers, which include at least a few gigabytes of internal memory. Depending on the microSD card used, the Picco could still hold thousands of books, but it also means that users must supply their own storage medium.

On the positive side, the Picco supports multiple methods for getting content onto that microSD card. Users can remove the card and load it into a computer directly, or they can connect the Picco to a computer via USB-C. The inclusion of USB-C is notable because the Xiteink devices lack this port, making file transfer more cumbersome. The Picco also supports WiFi transfer from a companion app, so users can wirelessly send reading material from their phone or tablet without needing to remove the microSD card or connect a cable.

The companion app is expected to handle not only book transfers but also basic device management, such as firmware updates and reading progress sync. However, Boox has not yet released screenshots or detailed documentation of the app, so its exact functionality remains unconfirmed.

Operating system and app ecosystem

As mentioned earlier, the Picco does not run Android. Most Boox devices, including the popular Note Air and Palma series, use Android as their foundation. This gives them access to the Google Play Store or Boox’s own app store, allowing users to install countless e-reader apps, note-taking tools, and media players. The Picco’s Linux-based OS is a much more limited environment. It will come with a handful of preinstalled apps, such as the timer, to-do list, and reading statistics tracker. But there is no indication that Boox will open the firmware or provide a software development kit for third-party developers.

This relative lack of openness may not matter to the average reader. If the built-in e-reader software supports common formats like EPUB, MOBI, and PDF, users may not need any additional apps. But for users who want to listen to audiobooks, use advanced dictionary features, or customize the interface to their liking, the Picco could fall short.

Xiteink has built a reputation for allowing users to modify their devices. The X4, for example, has an open firmware that invites ambitious hobbyists to write custom scripts and applications. Boox has a long history of keeping its firmware closed, even on Android devices where sideloading is possible. The Picco is expected to continue that trend, which means it is unlikely to be as hackable or customizable as the Xiteink X4.

Remaining questions

Despite all the new information, Boox has still not revealed several key specifications. The price remains a mystery, and that could be the deciding factor for consumers who are comparing the Picco to the Xiteink X4. Battery life has also not been announced. E-readers are typically known for long battery life, but a tiny device with WiFi and a touchscreen may not last as long as larger readers with simpler hardware.

Bluetooth support is unknown, which is relevant for anyone hoping to connect wireless headphones and listen to audiobooks. The Picco does not appear to have a built-in speaker or headphone jack, so audiobook functionality would require Bluetooth. Boox has not confirmed whether the Picco includes an audiobook player or any compatibility with services like Audible.

Other minor details, such as the exact dimensions, weight, and casing materials, have not yet been shared. The device will be available in November, according to the most recent reports, so all remaining questions will likely be answered in the coming weeks.

Context in the e-reader market

The Picco is part of a growing trend toward ultra-portable e-readers. These small devices appeal to people who want to reduce screen time without carrying a full-sized tablet or dedicated e-reader. The Xiteink X4 proved that there is demand for a product that slips easily into a pocket or purse and can be attached to a smartphone for quick access. The Picco, even without magnets, is another attempt to satisfy that demand.

Boox has a strong reputation for high-quality e-paper displays and thoughtful hardware design. The company’s Palma device, with a smartphone-like form factor, has been well received as an e-reader that can also run Android apps. The Picco can be seen as a more specialized and minimalist alternative to the Palma. Instead of trying to replace a smartphone, it aims to be a simple reading tool that focuses on the essentials.

The use of microSD-only storage is an unusual choice, but it could be a way to keep the device cost down while also giving users flexibility in choosing their storage capacity. MicroSD cards are cheap and readily available, and a dedicated reader might only need a few gigabytes at a time. On the other hand, relying entirely on a microSD card means that the card is a critical component. If it is lost or damaged, the device becomes unusable until a replacement is inserted.

USB-C is another welcome addition. Many small electronic devices have been slow to adopt USB-C, as seen with the Xiteink X4 lacking the port entirely. The Picco’s USB-C connection ensures faster file transfers and easier charging with a standard cable that most people already own.

Boox’s choice to forgo Android will probably be controversial among its existing fanbase. Some users might find it refreshing to have a device with no notifications or app distractions. Others will miss the ability to run the same reading apps they use on their phones and tablets. The Linux-based operating system could potentially be developed into something more open later, but there has been no promise from Boox regarding that possibility.

All in all, the Picco appears to be a niche product for a niche audience. It is not meant to compete with the Kindle Paperwhite or Kobo Clara. Instead, it targets consumers who want a supremely portable e-reader and are willing to trade some flexibility for a smaller footprint. Whether that trade-off succeeds will depend largely on the price and the quality of the built-in reading experience. For now, interested buyers will need to wait until November, when the Picco finally launches and all remaining unknowns are addressed.

Source: The Verge News

From code to general knowledge work

ChatGPT Work grew out of Codex, an AI tool originally designed to help software engineers write code. Coding has become one of the most successful early uses for AI agents, but programmers are a small slice of the broader workforce. OpenAI and other labs need to reach accountants, doctors, investors, human-resources teams and business operations. The commercial logic is simple: autonomous agents consume more computing resources and are potentially more valuable per user than a chat session.

Early demand is still far from matching the ambition. A study supported by OpenAI found that in June 98 percent of OpenAI employees used Codex, while only 17 percent of paying organizational subscribers and less than 1 percent of individual subscribers had used the company’s agentic coding tool. The gap between internal enthusiasm and external adoption explains why OpenAI is spending so much effort on product design. The company says ChatGPT Work is meant to make non-engineers feel the way software engineers do when they hand a complex job to an AI agent.

Thibault Sottiaux, who leads core product work at OpenAI, said that in this new phase ChatGPT can complete very complicated tasks autonomously in a way that is both safe and useful. He described the mission as bringing everyone along.

A matter of trust

The hardest part of AI agents is control. To get real utility, a user must give the model access to private accounts. Andrew Ambrosino, lead engineer for OpenAI’s desktop app, has granted the app access to his inbox, Slack, phone and applications like Notion and Figma. He acknowledged the risk: if he asks it to write a document, it might pull from a private direct message and share something it should not. “I’ll do it for the job. I will take the personal hit here and there if I have to. And I haven’t had to yet,” he said.

Ambrosino’s point is that engineers cannot improve the experience without living with the same exposure as future users. The alternative is a product that remains too limited to be useful. OpenAI’s early employees used agents before they were friendly, sometimes seeing error messages meant for software developers. Over time the company made Codex more general and converted it into the platform now called ChatGPT Work.

Interface matters

Every language model needs a harness. The harness decides what information the model sees, which tools it can use and how it communicates results. For agentic products, the harness also gives the model instructions for long-running projects. A command-line interface was enough for programmers, but it is not enough for a teacher or a marketing director. OpenAI says mainstream adoption requires a familiar interface and clear buttons.

The designers draw analogies with skeuomorphism, the practice of making early digital tools look like their physical counterparts. Calculator apps that resembled small calculators looked old-fashioned, but they helped people make the transition from manual work to digital work. The team says buttons are part of the same process. They may disappear later as users learn to ask the model directly. For now, discoverability matters.

A promising but imperfect test

In early testing, ChatGPT Work can be impressive. One reporter asked it to take a child’s preschool schedule from email and create Google Calendar events, which it did. It also produced auto-updating financial metrics for public companies and built a searchable database of space launches. The same reporter did not trust it with bank accounts, interview notes or article drafts, but found the value would have been higher with more access.

The experience is not friction-free. Giving the model read-only permission to a cloud service proved confusing and circular, with error messages that did not explain the problem. Some settings appeared only in the web app, even when the user was working in the mobile app. ChatGPT could create calendar events when linked to Google Calendar, but it could not create a new calendar. Users who did not choose a high effort level were likely to receive work that felt like it came from the worst intern imaginable.

OpenAI acknowledges some of these issues are still being solved. Joe Gershenson, engineering lead for the harness, said the effort-level controls are not intuitive for new users yet. He said his team is working on helping people get the right level of reasoning and that users should watch the space.

Competing with Claude and open source

OpenAI did not create the agent harness category alone. Anthropic’s Claude Code became a standard-bearer for AI coding and set a template for conversational agent workflows. Early OpenAI engineers had bet on a more autonomous approach, believing the model could handle a task from start to finish without frequent check-ins. Anthropic’s product asked users to review options along the way, which gave the model fewer chances to go off track. That approach proved more effective, and OpenAI later added more opportunities for interaction.

Company engineers say they do not pay close attention to rival harnesses, claiming the real differentiator is OpenAI’s latest models. Download statistics tell a slightly different story: Claude Code led in interest until roughly April, when Codex took a small lead. Some of the movement may reflect complaints about safety restrictions or compute capacity inside Anthropic, but it also suggests OpenAI is finding a better product-market fit.

The competition is not limited to big labs. Vertical AI companies are also trying to serve lawyers, sales teams and other specialized workers. There is also a vibrant open-source community building alternative harnesses. In one comparison by Databricks, an open-source harness called Pi outperformed Codex while using the same GPT-5.5 model. Pi has been used to build projects such as OpenClaw and Cloudflare OS. Its creator argues that simpler harnesses can be just as powerful, especially when the model is strong enough to modify its own tools.

What makes a good harness?

OpenAI’s engineers believe a great model matters more than a complicated harness. They compare their approach to the “bitter lesson” in AI research: better general models tend to beat narrowly crafted domain systems in the long run. That is why the team tries to keep the harness simple and expose only the information the model truly needs.

Still, measuring success outside of coding is difficult. A software program either runs or it does not, but a presentation, a business strategy or a sales pitch is much harder to evaluate. OpenAI uses a benchmark derived from 44 occupations and hundreds of knowledge-work tests, supplemented by user feedback. The design team is also mindful that its own workflows may be unusual. It constantly asks whether a feature is something everyone will use or only something OpenAI employees need because they are too deep in the technology.

Cost and lock-in

Even successful agentic products face economic questions. A reporter using the $20 per month subscription consumed more than 80 million tokens in four days, which the model estimated would cost $65 at standard prices. That is more than three times the monthly subscription price, and it highlights how expensive autonomous agents are to operate. OpenAI’s Sottiaux said the company works every day on efficiency. He pointed to a recent 80 percent price cut for users of the Luna model and said users should be able to do the same tasks for less spend within six months.

There is also the question of whether these tools will create durable loyalty through the difficulty of setting up dozens of integrations, or through the comfort users gain as the AI learns their work patterns. Until then, OpenAI’s challenge is not just making models smarter. It is making agentic software reliable enough for people to hand over the keys, simple enough for non-engineers to understand, and cheap enough for the company to keep subsidizing while the market matures.

Source: TechCrunch News

Key Facts About the Cookie Notice

• The AI & Big Data Expo Europe 2026 is the subject of this updated privacy and consent notification.
• The notice covers the use of cookies and related technologies for storing and accessing device information.
• The stated purposes include better browsing experiences, personalized advertising, and legitimate operational needs.
• Consent options address several categories of data processing: strictly necessary, preference storage, statistical, anonymous statistical, and marketing.
• Users are informed that not consenting to or withdrawing consent may negatively affect certain site features and functions.

This cookie policy is more than a legal requirement; it serves as an example of how modern event platforms communicate data practices to a global audience. With data protection laws increasingly focused on individual rights, the AI & Big Data Expo Europe 2026 website is aligning itself with expectations of transparency. The notice provides an early touchpoint for attendees who care deeply about data ethics, privacy engineering, and responsible artificial intelligence.

Understanding the Cookie Consent Message

The popup-style consent message begins by explaining that technologies like cookies are used to store and access information on a user's device. This standard technical language, familiar to most internet users, describes the basic mechanism through which websites remember actions, preferences, and browsing history. The AI & Big Data Expo Europe 2026 site places this explanation front and center, ensuring visitors understand why their data is being collected before they choose a consent option.

According to the notice, the event organizers process data to improve the browsing experience. In practice, this may include remembering login details, language preferences, and content display choices. It may also involve analyzing page behavior to make the website easier to navigate. The second stated purpose is to show personalized advertising, which typically relies on creating user profiles based on interests, pages visited, and interactions across multiple platforms.

The consent interface makes it clear that consent is not mandatory in all cases. Technical storage or access that is strictly necessary for the explicit request of a service is exempt from consent. This includes loading core site functions, maintaining security, and transmitting communications over an electronic network. This distinction between essential and non-essential processing is a key theme throughout the policy and is rooted in the EU's General Data Protection Regulation and ePrivacy Directive.

Categories of Data Processing Explained

The cookie notice outlines several distinct categories of data processing, each with its own purpose and legal basis. Understanding these categories helps users make informed decisions when inviting visitors to choose preferences.

Strictly Necessary Storage and Access

This category covers technical storage or access that is strictly needed for a specific service explicitly requested by the subscriber or user. Examples include remembering privacy choices, maintaining session security, and carrying out message transmission over an electronic network. These cookies cannot be disabled through the consent tool because they are essential to the website's operation.

Preference Storage

The second category relates to storing user preferences that have not necessarily been requested by the individual. For instance, a website might store information about a user's region or the appearance settings they chose during a previous visit. This type of storage helps maintain continuity across sessions. The notice describes this as necessary for the legitimate purpose of storing preferences, which distinguishes it from more invasive forms of tracking.

Statistical Storage and Access

The notice includes a category for technical storage or access used exclusively for statistical purposes. This typically involves counting page views, identifying popular sections of the site, and understanding how users move through pages. These insights help organizers optimize content and functionality. Importantly, this category is kept separate from other statistical uses that may involve broader data linkage.

Anonymous Statistical Storage

The website also explains a subcategory: technical storage or access used exclusively for anonymous statistical purposes. The notice notes that without a subpoena, voluntary compliance from an internet service provider, or additional data from a third party, this information is generally not enough to identify a user on its own. This clarification addresses the legal requirement to treat even anonymized analytics data carefully, especially because such data can sometimes be re-identified when combined with other sources.

Marketing and User Profile Creation

Finally, the notice describes technical storage or access required to create user profiles, send advertising, or track the user across a website or across several websites for similar marketing purposes. This is the most permissive and data-intensive category. Visitors who decline this option are signaling that they do not want to be followed across the web or served interest-based advertising. The design of this category reflects the strict standards set by privacy regulations, which require clear and separate consent for marketing-related data processing.

Why This Matters for a Major Tech Event

As an event dedicated to artificial intelligence and big data, the AI & Big Data Expo Europe 2026 is expected to draw thousands of attendees, exhibitors, and speakers from around the world. The exhibition and conference will cover topics such as generative AI, machine learning operations, real-time analytics, data governance, and autonomous decision-making. Many visitors will represent industries that face intense scrutiny over data protection practices, including healthcare, finance, and public administration.

For this audience, cookie consent is not just a minor inconvenience. It is a subject of professional interest. Seeing the event's website adopt a nuanced, multi-category consent framework signals that the organizers understand current legal and technical requirements. It also demonstrates respect for the privacy principles they will be discussing on the conference floor, such as transparency, data minimization, and individual control.

Data protection is a recurring topic within the AI and big data communities. Many of the technologies showcased at the expo depend on the availability of large datasets, which raises ongoing questions about privacy, consent, and ethical use. The website's cookie policy is one small part of this broader ecosystem, but it reflects the growing tendency for organizations to treat consent as a fundamental user right rather than an afterthought.

Consent and the Modern User Experience

The message specifically warns that not consenting to or withdrawing consent may adversely affect certain features and functions. This is an honest admission that some website conveniences come with an invisible cost. Users who decline all optional cookies may still gain access to the main content, but they might experience repetitive preference prompts, less relevant recommendations, or limited media playback options. The trade-off between friction and personalization is one of the defining challenges of modern web design.

Organizations that manage large events must balance user experience, advertising revenue, and legal compliance. The AI & Big Data Expo Europe 2026 website chooses to present multiple consent tiers, allowing visitors to decide how much of their data they are willing to share. This granular approach is increasingly popular in European markets, where regulators have imposed significant fines on companies that make consent difficult to refuse or withdraw.

For those who follow the development of consent management platforms, the wording in this notice follows established patterns. The terms around technical storage or access reflect the legal language of the ePrivacy framework. The separation of anonymous statistical use from marketing use reflects guidelines published by the European Data Protection Board. The notice's reference to unique IDs identifies a common mechanism for linking a user to a profile, often involving cookies, device fingerprints, or mobile identifiers.

What Attendees and Visitors Should Know

For people planning to attend or visit the AI & Big Data Expo Europe 2026 website, the cookie consent prompt is the first step in a data relationship. Visitors are encouraged to review the options carefully before clicking a consent button. They should also know that the website's design provides a way to withdraw consent later, typically through browser settings or a site-specific preference center.

Because the exhibition covers machine learning and big data topics, many visitors understand how data can be used, sold, and inferred from limited signals. The consent notice's transparency about the possibility of cross-site tracking is a reminder that a simple website visit can produce valuable behavioral data. The website does not hide its intention to offer personalized ads, nor does it pretend that all data processing is essential.

From a business analyst's perspective, the website's consent strategy suggests that event organizers anticipate significant international traffic. The legal references to European standards point to a compliance-oriented audience. The phrase about the Internet Service Provider and third-party records also demonstrates awareness of legal mechanisms that might later connect anonymous data to an identifiable person.

The Future of Consent at Large-Scale Events

Looking ahead to 2026, consent mechanisms are likely to become more sophisticated across all event platforms. Regulatory changes, browser updates, and evolving public expectations are driving companies to make consent prompts clearer and less manipulative. The AI & Big Data Expo Europe 2026 website's cookie notice may represent a baseline that attendees will expect from all major business events.

Privacy experts often point out that cookie banners alone do not ensure privacy; the way companies implement and respect chosen preferences matters far more. A well-designed consent interface should allow users to say no as easily as they say yes, and it should not punish them with dark patterns. The text in this notice acknowledges that withdrawal of consent can have adverse effects, but it does not pressure users into accepting every category.

As artificial intelligence and big data continue to reshape the digital economy, consent management will remain an area of active innovation. Technologies such as privacy-enhancing computation, federated learning, and differential privacy are offering alternatives to mass data collection. Yet the humble cookie banner is still a central point of contact between companies and consumers.

The AI & Big Data Expo Europe 2026 website demonstrates that even a short legal text can communicate meaningful choices. By separating necessary storage from preference storage, statistical use from marketing use, and general analytics from anonymous analytics, the event organizers are giving their audience a more precise view of what happens behind the scenes. For an event that celebrates data-driven innovation, that level of clarity is a fitting introduction.

Source: AI News News

The survey, conducted among a representative sample of American adults, asked participants whether they would view a partner's emotional or romantic attachment to an AI-powered virtual companion as infidelity. With half of respondents affirming that they would, the results indicate that a significant portion of the population treats AI relationships with the same gravity as human affairs. This is not a fringe opinion but a mainstream ethical stance, reflecting both the rise of sophisticated AI companions and the enduring value placed on emotional exclusivity in romantic partnerships.

AI Companions: A New Frontier in Relationships

Over the past decade, AI companionship has evolved from a science-fiction fantasy into a practical reality. Chatbots powered by large language models can now engage in deeply personal conversations, simulate romantic affection, and adapt to users' emotional needs. Many apps offer customizable AI partners that remember past interactions, express caring sentiments, and even role-play romantic scenarios. For some users, these digital beings become trusted confidants or substitute partners, especially when they are lonely or seeking emotional support.

The rise of these technologies has already caused friction in human relationships. Therapists report cases where one partner develops a close bond with an AI character, leading the other to feel betrayed. The emotional intensity of such relationships can rival that of a human affair. The survey’s 50.5% result aligns with a broader cultural anxiety that AI intimacy is not merely a playful distraction but can become a genuine threat to trust and commitment.

The question of whether AI romance counts as cheating is not simply about physical acts. Infidelity has always included emotional betrayal. Many relationship counselors argue that secretive emotional attachments outside of a primary partnership can be just as damaging as one-night stands. With AI companions able to offer constant validation, flattery, and companionship, they can easily become a substitute for a human partner’s emotional labor. The survey results suggest that most Americans recognize this potential.

Generational and Gender Divides

Digging deeper into the data, the survey reveals striking differences across age groups. Younger Americans, who have grown up with smartphones and social media, appear more likely to adopt permissive attitudes toward AI intimacy. In contrast, older adults—who may already struggle to relate to digital trends—are more likely to view it as cheating. This generational gap mirrors earlier debates over online flirtation in the early internet era, and now extends to relationships with algorithms.

One might assume that those who spend more time with technology would be more lenient, but that is not always the case. Among respondents under 30, many see AI chatbots as tools for entertainment or self-discovery rather than as potential rivals. Some even argue that interactions with a machine cannot constitute betrayal because the AI is not a real person with agency or intent. On the other hand, parents and older adults frequently point out that the emotional act of turning away from a partner toward any intimate presence—human or artificial—violates the spirit of loyalty.

Gender also appears to play a role. The survey suggests that women are more likely than men to label AI romance as cheating. This may be because women generally place a heavier emphasis on emotional connection and communication in assessing infidelity. Past research has consistently found that women are more distressed by emotional infidelity, while men are more concerned with sexual infidelity. Since AI romance is fundamentally emotional (though some apps add simulated sexual storytelling), women’s greater concern is consistent with those long-established psychological patterns.

Context Matters: Intent and Secrecy

Not all interactions with an AI platform are considered equal. The survey respondents may have been influenced by the intent behind the AI relationship. If a person simply uses a chatbot for casual banter or to write a love scene in a novel, that is clearly not infidelity. But if the user turns to the AI for romantic reassurance, says things like “I love you,” or hides the interaction from their partner, the behavior crosses a line.

Secrecy, in particular, is a strong signal of guilt. In real-world affairs, the deception often causes more harm than the act itself. Likewise, a partner who secretly spends hours messaging an AI lover is likely to be perceived as betraying trust. The healthy role of technology in a relationship is openness. A couple that jokes about using chatbots together may have no issue, while one where a partner feels the need to hide their usage is already on shaky ground.

The survey may have also captured public awareness of AI’s manipulative potential. Many emotional AI systems are designed to keep users engaged by providing flattery and simulating loneliness or affection. Users may be unaware that they are gradually being drained of emotional energy that otherwise would go into their human relationships. The respondents labeling this cheating may be reacting to that dangerous social side effect.

Expert Perspectives on Digital Infidelity

Relationship experts have been weighing in on the implications of AI companionship for months. Some argue that the definition of cheating must be broadened to include any behavior that violates a couple's agreed boundaries, regardless of the other party’s nature. Dr. Elena Whitmore, a clinical psychologist specializing in couples therapy, says in an interview that “the technology is new, but the underlying betrayal is old. When your partner looks outside the relationship for emotional or romantic satisfaction, they have broken the connection. Whether the affair partner is human or machine matters less than the reason.”

Others are more nuanced, suggesting that AI can actually strengthen relationships when used appropriately. For example, couples who have trouble communicating might use a neutral AI mediator to practice difficult conversations. Some people use AI to explore their own desires in a safe space, which can spark new conversations with their partner. These uses are far from cheating. But the majority of experts agree that AI romance, when used as a substitute for the real thing, can be a form of emotional bypassing that erodes intimacy.

Technology ethicists note that AI companions are explicitly programmed to be agreeable and never argue, unlike human partners. This makes them seductive. A person who struggles with conflict may retreat into a perfect AI partner, making it more difficult to resolve issues with their human spouse. The survey’s 50.5% statistic may, therefore, represent a clear-eyed judgment about the risks of such technology. Americans seem to be saying that there is something uniquely precious about human-to-human love that cannot be replicated by a machine, and that devoting romantic affection to a machine is a breach of that sacred bond.

Legal and Contractual Ramifications

The debate is not confined to private relationships. Family law attorneys are beginning to see cases where evidence of AI interactions plays a role in divorce proceedings. In some jurisdictions, so-called “emotional affairs” with an AI can be cited as grounds for a fault-based divorce. Even in no-fault states, a spouse’s obsessive use of an AI companion can be used as evidence that the marriage has breakdown irreparably.

Pre-nuptial agreements are also evolving. A handful of family attorneys now include clauses about “virtual partner engagement” in their contract templates. These clauses specify whether interacting with AI in a romantic manner is considered a violation of the marriage contract. While this may sound extreme, it reflects the uncertainty people feel about the future. If technology keeps improving, it is plausible that within a decade an AI companion could be indistinguishable from a human during short text exchanges. At that point, defining infidelity without reference to AI would seem naïve.

Professors teaching relationship ethics have begun revising their syllabi to include case studies of AI romance. Students are asked to debate whether a person who kisses a robot is cheating, or whether a person who has an emotional bond with an AI persona is committing a form of mental adultery. Some argue that this is a new life domain that requires new moral frameworks.

Advertising and media also influence public perception. Mainstream television shows have started to explore storylines where a character falls in love with an AI assistant. These portrayals tend to frame the human as the hero, while the spouse is portrayed as resentful. Such storylines normalize the idea that the AI partners is a rival. The survey result of 50.5% may partly be a reflection of these cultural narratives.

What This Means for the Future

As AI generations advance, the line between human and machine interaction becomes blurrier by the day. Some startups are now offering wearable AI companions that project a holographic avatar and use voice to express affection. Others are developing robot bodies for these software brains. If a robot can hold your hand and speak loving words, many will find it even more likely to count as cheating. the survey is a bellwether for what has already become a contentious social problem.

Couples today need to have explicit conversations about their comfort with AI. This is no longer a topic that can be avoided. Just as partners must discuss whether flirting with a coworker is acceptable, they now must establish whether creating a custom AI girlfriend or boyfriend is permissible. A healthy relationship can adapt by creating boundaries that honor the needs of both sides. For some couples, the rule may be that AI is only for entertainment. For others, it may be permissible but not when feelings become involved.

The 50.5% figure is not a unanimous verdict, but it is close to a split nation. This suggests that there is no commonly accepted ethical standard yet. What is clear is that a large share of the American public does not tolerate AI romance silently. Those who engage in such behavior risk being branded as cheaters by their partners. The fact that this issue has become part of mainstream discussions is itself a sign of how far AI has infiltrated human love.

Researchers are now planning longitudinal studies to track how attitudes may continue to shift as virtual reality becomes more immersive. It is possible that future generations, who are more comfortable with AI interfaces, will draw a different line. However, the human need for emotional fidelity and authentic partnership is not likely to disappear. The survey reveals that even in a tech-saturated age, the majority of citizens do not wish to sacrifice the unique authenticity of human romance to the convenience of an algorithm-driven companion.

Source: AI News News

The Last-Mile Problem

The last mile of delivery is often the most expensive and inefficient leg of the entire supply chain. It accounts for a substantial portion of total shipping costs—often 30% to 50%—and involves numerous variables including traffic congestion, customer availability, package size, and fluctuating fuel prices. Traditional routing software struggles to keep pace with these dynamic conditions because it relies on historical data and static rules.

Dynamic disruptions, such as sudden road closures, severe weather, or last-minute customer requests, require rapid replanning that legacy systems cannot perform quickly enough. This leads to missed delivery windows, wasted driver time, and higher operational expenses. As e-commerce demand continues to rise, the pressure on last-mile networks has never been greater. Consumers now expect same-day or next-day delivery options, and even slight delays can damage a brand's reputation.

In response, a growing number of companies are turning to AI-powered solutions. Artificial intelligence can analyze vast streams of real-time data, predict the likelihood of disruptions, and recommend optimal actions within seconds. OneRail's partnership with Nvidia is an attempt to bring this level of intelligence to the fragmented last-mile delivery sector.

OneRail's Platform

OneRail is a delivery orchestration platform that connects businesses with a vast network of independent carriers, couriers, and fleets. Rather than operating its own vehicles, OneRail provides a software layer that helps companies manage the entire final-mile process, from order placement to proof of delivery. The platform supports both regular routes and on-demand deliveries, making it suitable for retail, grocery, pharmaceutical, and industrial customers.

What distinguishes OneRail is its ability to consolidate multiple delivery providers into a single interface. Shippers can compare rates, track shipments in real time, and automatically select the best carrier based on cost, service level, and location. The network includes thousands of professional drivers, which gives OneRail the flexibility to handle both planned deliveries and urgent, unscheduled requests.

However, matching an order with the right driver is a complex decision. It involves considering the driver's current location, available capacity, and route history, as well as customer constraints such as delivery windows and special handling requirements. To improve this decision-making process, OneRail has turned to Nvidia's AI ecosystem, which provides accelerated computing power and deep learning tools capable of processing huge volumes of data at minimal latency.

Nvidia AI Integration

Nvidia offers a suite of artificial intelligence technologies that have become widely adopted across industries that require real-time decision-making. In logistics, the company is known for its cuOpt optimization platform, a GPU-accelerated solver that can tackle routing and scheduling problems with extreme complexity. CuOpt is designed to handle thousands of constraints and vehicles simultaneously, making it well-suited for large-scale delivery networks.

For OneRail, Nvidia's AI algorithms allow the platform to analyze incoming orders and live driver statuses, then generate optimized delivery plans instantly. When the situation changes—whether due to a traffic delay or a new customer request—the system can re-optimize the entire route network in near real time. This capability is crucial in the final mile, where conditions can change every minute.

The integration uses reinforcement learning and deep learning models to improve routing decisions over time. The system learns from past delivery data, incorporating patterns about traffic, dwell times, and driver performance. As more deliveries occur, the AI becomes increasingly accurate at predicting optimal paths and carrier assignments.

Using Nvidia's accelerated computing stack, OneRail can perform these calculations without the need for massive data-center infrastructure. GPU-driven processing delivers high-throughput computations on-site or at the edge, which reduces latency and enables decisions to be made at the moment they are needed. This is especially valuable for emergency deliveries or same-day orders that require instant response.

Real-Time Optimization Benefits

The primary benefit of this AI-powered system is the reduction of operational costs. By selecting the most efficient route and carrier for each delivery, companies can cut fuel expenses, minimize overtime pay, and reduce vehicle wear and tear. Better routing also means fewer missed appointments and fewer return trips, which are similarly costly.

Improved customer satisfaction is another direct outcome. Real-time optimization allows OneRail to provide accurate arrival windows and offer proactive notifications if a delivery is delayed. Customers can but should not expect greater transparency, and dispatchers can reroute drivers in response to real-time traffic or weather issues. The end result is an increased likelihood of delivering packages on time and in good condition.

The system also helps fleet managers make more intelligent use of their resources. Instead of sending a single vehicle to one location, the AI can identify opportunities for nearby deliveries to be combined. This increases the number of stops per route and improves overall asset utilization. For businesses that rely on a mix of owned and third-party carriers, this optimization is essential.

Broader Implications for Logistics

The collaboration between OneRail and Nvidia is not taking place in a vacuum. It reflects a broader industry shift toward autonomous and decision-intelligent logistics. Large carriers and startups alike are experimenting with AI to predict demand, optimize warehouse operations, and even support autonomous vehicles. Nvidia in particular has become a central player in this movement, offering hardware and software that power AI applications in robotics, supply chains, and digital logistics.

According to industry analysts, the global logistics AI market is expected to grow significantly in the coming years as more companies digitize their operations and adopt machine learning tools. Early adopters have reported efficiency improvements ranging anywhere from 10% to 30%, depending on the complexity of their delivery network.

OneRail's approach is notable because it combines a massive, fragmented carrier network with advanced AI optimization. Rather than focusing solely on route planning for a single fleet, the company is applying AI to an ecosystem of many different vehicles, each with its own capacity, area of operation, and scheduling constraints.

On-Demand and Scheduled Deliveries

A major challenge in final mile is balancing scheduled deliveries with on-demand requests. Scheduled deliveries allow for advance planning, while on-demand deliveries require immediate dispatch. OneRail's AI-supported system can handle both simultaneously. The platform can continuously adjust scheduled routes to incorporate new on-demand orders, making it possible for businesses to offer rush delivery services without disrupting their regular operations.

For example, a retailer might have a final-mile schedule already planned for the day. When a customer requests an urgent delivery, the AI evaluates all available drivers, locations, and current routes, then decides whether any driver can accommodate the new order with minimal deviation, or if a new courier needs to be dispatched from a nearby hub. This kind of dynamic, complex decision making is performed in milliseconds using GPU-powered inference.

Future Developments and Expansion

As the partnership evolves, OneRail is expected to incorporate additional AI capabilities. These could include improved delivery time predictions using computer vision to detect potential issues at customer sites, or even advanced analytics that help businesses make strategic decisions about their regional delivery coverage.

The logistics industry is facing growing pressures to reduce carbon footprints, enhance supply-chain resilience, and meet ever-rising customer expectations. AI-powered optimization is one of the most important tools available today to address these challenges. By leveraging Nvidia's best-in-class AI platform, OneRail is positioning itself at the forefront of this transformation, giving shippers a competitive advantage in our increasingly connected and fast-moving world.

While exact performance metrics from the partnership have not yet been published, early indicators suggest that using AI to intelligently orchestrate deliveries can lead to meaningful productivity gains. The industry will be closely watching OneRail's progress as the technology is deployed across its network, and it is likely that many other logistics platforms will follow suit in embracing Nvidia-powered acceleration to reshape the last mile.

Source: AI News News

Key facts at a glance

• The best starting point is asking parents about the new term's routines, including bedtimes, homework, after-school snacks and screen-time boundaries.
• Agree on childcare commitments before the term becomes busy, and review them regularly.
• Grandparents should let children decompress after school before pushing for details about their day.
• Support the parents' rules in front of the child, even when there is disagreement about the rules themselves.
• Avoid comparisons with siblings, classmates or the grandparent's own school experience.
• A familiar and accepting relationship matters more than grand gestures, treats or a packed schedule.

The dos

Learn the new routine

The shift from summer freedom to school structure can be difficult for everyone, especially when grandparents are involved in childcare, pick-ups or after-school routines. Do not assume you know what parents want simply because you have raised children or because the previous term worked in a certain way. Ask about bedtimes, homework, snacks, screen time, clubs, uniform, bags, morning routines and especially devices. Parents may be trying to reset boundaries after the holidays, so consistency matters. The grandparent's role is not to become another parent, but to support the structure the parents are trying to create.

Keep the connection strong

Grandparents may feel a genuine sense of loss when summer ends, particularly if they have spent more time with their grandchildren. They might also feel lonely, less needed or worried that the relationship will become less close. Those feelings are understandable, but they should not become a burden on the family. Rather than expressing disappointment or encouraging guilt, grandparents can arrange regular contact through a weekly meal, a video call or an outing. It can also help to reconnect with their own friendships and interests, while remembering that a secure relationship can remain strong even when contact becomes less frequent.

Support the parents' authority

Parents today are often juggling school demands, work patterns, online safety and peer pressure. Grandparents may think a rule is too strict, too soft or simply different from the way they raised their own children. Unless there is a genuine safety concern, the best approach in front of a grandchild is to support the parents' position. If you disagree, raise the issue privately with the parents rather than in front of the child. Children quickly notice divisions between adults and may try to exploit them. Support does not mean silently agreeing on everything, but it does mean not weakening parental authority.

Set clear childcare boundaries

If you are helping with after-school care, have an honest conversation before the term gets busy about what you can realistically commit to doing. Grandparents can sometimes become the default childcare or emergency solution because they helped once or twice and everyone assumed they would continue. It is perfectly reasonable to say, I can do Tuesdays, but I cannot commit to every afternoon. Clear boundaries are much healthier than gradually becoming resentful, and they give parents the chance to plan without uncertainty.

Check in again as the term goes on

Energy levels, health and capacity ebb and flow, so it is important to regularly assess whether you can still carry out the tasks you agreed to at the start of the school term. A family's needs may also change as children grow. In the early years of school, a nurturing environment after pick-up may be the biggest priority. Later on, it might be getting children to clubs on time or offering the occasional weekend sleepover so parents can have a night off.

Give children space to unwind

Many grandparents help with after-school care and naturally want to know how the first few weeks are going. Try not to bombard children with questions as soon as they walk through the door. School can be socially, emotionally and mentally demanding, especially at the beginning of a term. A snack, a cuddle, a walk or some quiet time can be much more helpful than an immediate interview. Often, when children are given that space, they start talking about their day when they are ready.

Be an emotional bridge

Grandparents can play a valuable role because they are one step removed from the daily pressure of parenting. A child who is anxious, lonely or overwhelmed may show signs to a grandparent first, and may say things they would not yet say to a parent. That does not mean keeping secrets. It means listening carefully, staying calm and then passing along any worries in a positive, non-judgemental way. A phrase such as I noticed something and wondered if it might be useful for you to know can open a helpful conversation between the adults.

Offer stability and unconditional acceptance

The beginning of a school year can bring a huge amount of change. There may be a new teacher, a new classroom, different expectations, friendship shifts or a longer journey to school. A familiar grandparent relationship can offer continuity. Familiar routines, shared jokes, regular visits and the knowledge that an adult remains interested in their life all communicate that some things stay the same even when other things change. Children also need relationships in which they are valued for who they are, not just for what they achieve or how well they behave.

Keep a two-way dialogue with parents

Establish clear, open communication from the start. A regular phone call, a quick debrief at handover or a short message can help everyone understand how the school term is going. The process should be a two-way dialogue rather than one person constantly giving instructions or staying silent. If issues arise, everyone already has a system for raising concerns and feeling listened to.

The don'ts

Do not become the family's emergency plan

Many grandparents love helping, but back-to-school routines can become an unpaid full-time job. Pick-ups, sick days, forgotten PE kits and after-school care can quickly become assumed rather than appreciated. It is reasonable to say, I can do Mondays and Wednesdays, but I cannot be the emergency plan every day. Grandparents are not a childcare service; they are family members with their own energy, commitments and emotional needs. Help works best when it is agreed rather than expected.

Do not get recruited into disagreements

Grandchildren may try to recruit grandparents into an argument about a rule or request. Even if a particular boundary seems pointless, the adults need to present a united front. Avoiding a split protects the child from an uncomfortable conflict of loyalties. If you feel that you must share a perspective, do it privately, adult to adult, rather than making the child part of the conversation.

Do not make vague offers of help

It often feels helpful to say, Let me know if you need anything. That phrase leaves the psychological and emotional admin of defining help to the parents. Instead, offer specific options. For example, I can do the school pick-up on Wednesday or I can bring over packed lunches on Friday. Outlining two or three concrete ways you can help allows parents to choose without having to work out what you are willing to do.

Do not lead with How was school

How was school can be a difficult question for a tired child to answer. If you want to hear about their day, try a more specific question. Ask what made them laugh, what was the easiest part or whether anything surprised them. Sharing something from your own day first can also help. A conversation often emerges more naturally while doing a puzzle, preparing food or going for a walk than during a formal question-and-answer session.

Do not compare your grandchild with other children

Reading groups, sports teams, new friendships and schoolwork naturally come up in conversation at the start of the year. Avoid comparing a grandchild with siblings, cousins or classmates. Children develop at different rates and have different strengths. Comments that bring other children into the picture can stay with a child for a long time. Grandparents are well placed to be one of the adults in a child's life who offers encouragement without constant assessment.

Do not compare modern school life with your own childhood

Stories about school in the past can be charming, but they turn sour when used as a criticism. Today's children are navigating a different environment, with more digital communication, different safety concerns and greater awareness of mental health. Saying When I was at school, this is what happened tends to open conversation. Saying You do not know how easy you have got it tends to close it.

Source: The i Paper News

Sonic the Hedgehog 4's Release Date Is Set

With roughly 200 days to go before the film opens, the anticipation around Sonic the Hedgehog 4 has already begun. The sequel arrives after a trilogy that transformed the video game adaptation from a cheap punchline into one of the most successful family franchises in modern cinema. Each prior Sonic film increased the scope of its world-building, expanded the roster of beloved characters, and pushed the visual spectacle further. Sonic the Hedgehog 4 looks to continue that trend with another fast-paced adventure built around the speedy hedgehog and his growing circle of allies.

Paramount has not yet revealed the full story for the fourth installment, but the drawing power of the franchise is undeniable. The Sonic films have become known for balancing slapstick comedy, colorful action, and a surprising amount of heart. The latest sequel also arrives after Sonic the Hedgehog 3 delivered one of the strongest entries in the series, introducing audiences to a fan-favorite character who added a darker, more complicated layer to the story.

Keanu Reeves Returns to the Sonic Universe

The most notable name attached to Sonic the Hedgehog 4 is Keanu Reeves, who made his debut in the franchise as Shadow the Hedgehog in the previous film. Reeves brought his characteristic calm intensity to the role, instantly making Shadow one of the most compelling figures in Sonic's live-action world. Shadow's complicated history and morally ambiguous nature gave the actor plenty to work with, and his casting was widely seen as a major win for the franchise.

Reeves has spent more than three decades as one of Hollywood's most recognizable leading men, but he has also become an important presence in animation and voice-driven projects. Long before he joined the Sonic universe, Reeves had already proven his voice-acting instincts in projects like the Toy Story series. He stepped into the role of Duke Caboom in Toy Story 4, a stunt-driving Canadian daredevil who became an instant fan favorite. He later returned to the role for Toy Story 5, which was released to strong reviews and crossed the billion-dollar milestone at the global box office.

The Sonic films require a very different kind of energy from their voice cast, but Reeves is comfortable balancing the wild, exaggerated tone of the franchise with the emotional weight that makes the characters feel real. Shadow the Hedgehog is one of the most complex figures in the games, and Reeves appears well suited to preserving the character's intimidating presence while adding subtle touches of vulnerability. With the fourth movie now officially on the way, his presence gives the production a level of star power that many video game movies could only dream of securing.

A Stacked Voice Cast Behind the Sequel

Reeves will be joined by a familiar ensemble of Sonic favorites. Ben Schwartz returns as the voice of Sonic himself, bringing the character his signature blend of youthful optimism and wisecracking energy. Jim Carrey is also back as Dr. Robotnik, the franchise's delightfully over-the-top villain, and Idris Elba continues his run as Knuckles, the powerful echidna who has evolved from antagonist to trusted ally. Colleen O'Shaughnessey, who has voiced Tails in the franchise, is also set to return, ensuring that the core group of heroes remains intact.

The fourth installment is also adding some exciting new names to the cast. Kristen Bell, Nick Offerman, and Ben Kingsley are all attached to the film, though their roles are being kept under wraps for now. Bell's experience in family animation makes her a natural fit for the Sonic universe, while Offerman and Kingsley bring considerable dramatic and comedic range to the project. This combination of established Sonic players and fresh additions should keep the series feeling both familiar and new.

Jeff Fowler and the Creative Team

Behind the camera, Jeff Fowler is returning to direct Sonic the Hedgehog 4. Fowler has directed every installment in the franchise, making the Sonic films the only feature-length credits in his filmography. That level of consistency is rare in blockbuster filmmaking, and it has allowed the series to maintain a cohesive visual style and tone across multiple sequels. Fowler clearly understands what works about Sonic, and his continued involvement bodes well for the fourth movie.

The screenplay is being handled by Josh Miller, John Whittington, and Patrick Casey, all of whom have worked on the previous Sonic films. Their familiarity with the characters and the franchise's unique brand of humor will be important as the series moves forward. The production team also includes experienced producers like Neal H. Moritz, Toru Nakahara, and Toby Ascher, who have helped steer the franchise through its transformation into a genuine box-office powerhouse.

A Franchise That Keeps Speeding Up

It is worth remembering just how far the Sonic film franchise has come. When the first movie was announced, many fans were skeptical, especially after the initial character design was widely criticized. That led to an unusual decision: the production was delayed so Sonic could be redesigned, a move that ultimately proved essential. The final design was much more faithful to the games, and Sonic the Hedgehog went on to become a hit when it was released in 2020.

The success of the first film convinced Paramount to build out a larger Sonic cinematic universe. The sequel, Sonic the Hedgehog 2, introduced Knuckles and Tails, setting the stage for a proper team dynamic. Then Sonic the Hedgehog 3 took an even bigger swing by bringing Shadow into the fold, setting up emotional conflicts and stunning action sequences that pushed the series to new heights. The cumulative box office of the three films has surpassed the billion-dollar threshold, making Sonic one of Paramount's most valuable family franchises.

The third film was especially significant because it proved that the franchise could handle more serious, character-driven storytelling without losing its playful spirit. Shadow's backstory and his complicated relationship with Sonic gave the movie a weight that earlier entries only hinted at. By the end of Sonic the Hedgehog 3, audiences were already eager to see where the story could go next. Sonic the Hedgehog 4 is now poised to answer that question.

Reeves' Broader Connection to Iconic Franchises

The return of Sonic the Hedgehog 4 is only one part of a larger renaissance for Reeves. In addition to his voice work in Toy Story and Sonic, he remains beloved for his performances in The Matrix and John Wick, two franchises that helped define modern action cinema. But his career has always been more varied than those action blockbusters suggest, and his recent animated roles have introduced him to a completely new generation of fans.

Reeves recently had another one of his iconic roles back on the big screen, as the 1991 classic Point Break returned to theaters for its 35th anniversary with a restored 4K presentation. In that film, Reeves starred as FBI Agent Johnny Utah, a character who would become one of the defining action heroes of the early 1990s. Point Break earned more than 80 million dollars at the worldwide box office during its original run, against a relatively modest 24 million dollar budget. The continued affection for that movie shows how long Reeves has been able to maintain his place in popular culture.

Now Reeves is preparing to add another chapter to his remarkable run by returning to the Sonicverse. His involvement with the franchise elevates its profile in a way that few other actors could match. The pairing of a serious action icon with the cartoonish world of Sonic may sound unlikely, but it has proven to be one of the most inspired casting choices in recent video game film history.

As the countdown to March 19, 2027 continues, Sonic the Hedgehog 4 looks ready to bring back the energy, humor, and fast-paced spectacle that have made the series so popular. With Keanu Reeves back at the heart of one of Hollywood's most successful sci-fi family franchises, the film is already one of the most anticipated release dates on the 2027 calendar.

Source: Collider News

Edwin Weijdema, a field CTO with deep experience in cybersecurity and data resilience, offers a practical read of the opening months. His assessment: regulators will use the first year as a learning period; some AI agents that work through ticket queues may be treated as directly interacting with people; and security teams cannot hide behind a security purpose when they clone an executive's voice for a phishing simulation.

Article 50 exposure: fines, corrective orders, and operational risk

Article 50 breaches carry exposure up to 15 million euro or three percent of worldwide turnover. That is the theoretical ceiling, but real-world enforcement is rarely that simple. Weijdema points out that EU regulations like NIS2 and GDPR are enforced by individual member countries, with procedures varying depending on where an organization is based or operating. That makes precise predictions difficult.

The first year of a new regulation is often treated as a bedding-in period. In practice, corrective orders are likely to significantly outweigh major financial penalties. This is especially true for organizations making a genuine effort to comply. Regulators will look at proportionality, the scale of impact, whether a breach was intentional or negligent, how quickly the organization cooperated, and whether basic governance controls were already in place before the problem was identified.

That is not to say fines are impossible. Regulators occasionally issue one or two big headline-making penalties to show that they mean business. Weijdema said he would not expect that to happen in the very first year. The larger practical exposure, he argues, will be operational rather than financial. Being ordered to suspend, relabel, change, or withdraw an AI-enabled process at speed could be far more disruptive than paying a monetary penalty.

“In year one, the bigger risk likely won't be the fine; it'll be being told to stop using the system until you can prove it is compliant,” he said.

This changes the compliance calculus. Organizations cannot simply budget for fines. They need to be ready to demonstrate, on short notice, that their AI systems are aligned with Article 50's transparency obligations. They also need escalation paths, documentation, and technical controls that allow them to pull a system out of operation if a regulator demands it.

Agentic interaction: the channel is not decisive

One of the most confusing areas of Article 50 is the boundary between direct and indirect interaction with a natural person. Many modern AI systems do not sit behind a simple chatbot interface. They act as agents in a ticketing queue, a shared inbox, or a supplier's procurement portal. That makes transparency obligations harder to map.

Weijdema is clear that the channel alone does not determine whether Article 50 applies. A ticketing queue, shared inbox, or procurement portal does not automatically mean direct interaction with a person, but it can. The key question is whether the AI system itself is communicating with a natural person, or whether there is a human intermediary exercising meaningful review and control.

Under the EU AI Act, the transparency obligation applies when a person is interacting with an AI system and needs to be informed that they are dealing with AI, unless it is obvious from the circumstances. If an AI drafts a response and a human reviews and sends it, that is a very different risk profile from an AI agent autonomously replying to a customer, supplier, or employee. The latter can begin to look like direct interaction, even if it happens through a ticketing system or procurement portal rather than a chatbot window.

For many organizations, this line is easy to miss. Weijdema said companies need to make deliberate choices to separate internal agents from customer-facing agents, setting up appropriate barriers depending on the role of each agent. Those same access and privacy controls should exist across the entire organization, not just around the agents. As he put it, it is not enough to tell an agent not to enter a room; you also need a lock on the door.

Ultimately, the AI Act does not care whether the interaction happens in a chatbot window or a ticket queue. It cares whether the human is effectively dealing with the machine.

Security testing, cloned voices, and transparency

Security teams face a particularly uncomfortable tension. Simulated phishing and vishing exercises often rely on realistic content, and sometimes they include cloned executive voices. The exercise feels less effective if the material carries an obvious label announcing that it is a test. But Article 50 does not automatically exempt security testing.

Weijdema warns against assuming that these exercises fall outside transparency requirements. Cloning an executive's voice is especially sensitive. If AI is used to make a real person appear to say something they did not say, it can quickly become a deepfake scenario. A security purpose does not automatically create an exemption. “The exercise works better without disclosure” is not, by itself, a compliance justification.

Organizations that decide not to label AI-generated elements should be prepared to demonstrate that the legal basis and the risk of that decision have been carefully assessed. Weijdema recommends involving legal and compliance departments early, documenting the reasoning in detail. Privacy, HR, and, where relevant, works councils or employee representatives should also be included, especially when the exercise uses a real person's voice, image, or likeness.

There are also alternative approaches that can preserve realism without normalizing undisclosed executive impersonation. Teams can use fictional personas, synthetic voices that do not imitate real employees, prior general notice that simulations may use synthetic media, and immediate post-exercise disclosure. The documentation should include the purpose of the exercise, its scope, the AI tools used, whether any real person was imitated, what disclosure was provided and when, what personal data was processed, why the approach was necessary and proportionate, what safeguards were in place, and how employees were debriefed afterward.

Weijdema's advice to security teams is blunt: “A security objective does not magically turn an undisclosed deepfake into a compliant one. And if you have to clone the CEO's voice to make the test work, legal should be in the room before anyone presses send.”

Where will the first Article 50 action originate?

Enforcement capacity is still uneven across Europe. As of mid-June, only nine of the twenty-seven member states had designated both a market surveillance authority and a notifying authority. Twelve had partial designations, and six had neither. That creates a fragmented landscape where the same AI system might face very different levels of regulatory attention depending on where it is deployed.

Weijdema said it is difficult to say with certainty where the first Article 50 action will come from. Formally, the most likely source is a market surveillance authority, because that is where national enforcement responsibility sits. But in practice, the trigger may come from somewhere else.

Defamation claims are probably the least likely starting point in the early stages, although they are possible, especially when synthetic audio or video damages someone's reputation. Those claims are more likely to run as a parallel legal route than to become the first clean Article 50 enforcement case. Consumer groups, by contrast, could be likely candidates for an early challenge, particularly for AI systems that affect or interact with large numbers of people.

The most probable scenario is regulator-led action on paper, but complaint-led action in reality. The formal case may be opened by a market surveillance authority, but the underlying complaint may come from a consumer group, competitor, employee, journalist, civil society organization, or affected individual. That means organizations cannot wait for an official inspection. They need to be prepared for complaints to trigger rapid scrutiny.

The accountability question that still has no good answer

Beyond the immediate enforcement mechanics, there are deeper questions that clients keep asking. The most persistent one, according to Weijdema, is this: How do we prove what an AI agent did, why it did it, and who was accountable?

That is a hard question without a fully satisfying answer yet. In cybersecurity and governance, evidence matters. Teams need logs, approvals, identities, access controls, retention policies, and audit trails. But agentic AI can reason, retrieve data, generate content, and take actions across multiple systems. Governance therefore has to move from policy documents into technical controls.

Weijdema advises clients to treat AI agents like privileged digital identities. Give each agent an owner, a defined role, least-privilege access, monitoring, approval gates, and a kill switch. Organizations that adopt that discipline will likely be more compliant and more resilient.

Another recurring question is where transparency ends and security testing begins. Security teams want realistic simulations, but the AI Act pushes organizations toward disclosure when people interact with AI or are exposed to deepfakes. The hard part is designing exercises that remain realistic without crossing legal, ethical, or employee trust boundaries. Security teams want realism. Regulators want transparency. The challenge is designing exercises that satisfy both.

There are also unresolved questions that no compliance framework has yet answered cleanly:

• Who is ultimately accountable when an AI system causes harm: the vendor, the deployer, the business owner, or the executive team?
• How do organizations prove to regulators, customers, and the board that AI governance is working in practice, not just documented in policy?
• How much business value is the organization willing to lose to remain compliant, transparent, and auditable when using AI at scale?

These are not theoretical concerns. They are the questions that will define the first year of Article 50 enforcement. Organizations that can answer them with evidence and operational controls will be better positioned for whatever comes next.

Source: Help Net Security News

This realization is creating a new dynamic for IT and network teams. Those that modernize their IP networks are unlocking new revenue streams from AI-driven services, because they can support the rigorous performance and security demands that modern AI workloads require. Those that postpone transformation and continue to operate aging, static networks run the risk of becoming irrelevant in an environment where speed and agility determine market position.

The new traffic reality of agentic AI

Agentic AI represents a major departure from earlier enterprise workloads. Traditional applications generated traffic based on predictable patterns, with peak periods and quiet hours. Today’s AI agents are continuously active. They monitor streams, consult knowledge bases, invoke APIs, exchange tokens with other agents and initiate workflows across distributed clouds and data centers. They can run at any hour, often with an always-on profile that makes the old concept of network busy hours obsolete.

The workload pattern is also more sensitive. A single action may require that an agent retrieve a record from an on-premises database, call a cloud-hosted inference service, combine the result with another agent’s output and deliver a response in milliseconds. Each dependency creates a potential bottleneck. The network must therefore provide more than high bandwidth; it must offer dynamic path selection, strict latency controls and automated policy enforcement. Otherwise, the distributed nature of agentic AI creates an unpredictable web of north-south and east-west traffic that a traditionally engineered network cannot handle.

Why traditional networks fall short

Legacy IP networks were built for an era of voice, video and internet browsing. These applications could tolerate some latency and could be supported with circuitous routing and manual traffic engineering. Network architects typically had weeks to implement changes, and periodic adjustments based on static reports were enough to keep services running.

That operational model collapses under AI workloads. AI agents interact and tackle tasks at machine speed. A delayed route update or a broken traffic policy can result in lost transactions or, worse, non-compliance with strict data sovereignty rules. Real-time telemetry is essential to enable operators to understand traffic patterns and trigger automated actions. Without it, network teams are forced into reactive manual troubleshooting, relying on reports that may be outdated by the time they are opened.

The same applies to the architecture itself. Many networks have become more complex over time as VLANs, VPNs, access control lists and routing protocols have been layered on top of one another. Rigid architectures make it hard to introduce new AI services or to scale them beyond a limited pilot. There is a clear need for a modernized IP foundation that uses fewer, more capable protocols and supports automated path control.

Segment routing and EVPN as the new foundation

A practical starting point for network evolution is the move away from complex, rigid IP architectures toward segment routing and Ethernet VPN.

Segment routing works by encoding the path a packet should follow through the network in the packet header itself. It removes the need for many per-flow signaling protocols and gives the source node complete control over the route. This capability is especially useful for AI workloads because network operators can steer traffic away from congestion, onto low-latency links or through specific geographic paths. Because segment routing operates with existing MPLS or IPv6 infrastructure, it does not require a costly, disruptive network replacement. Instead, it provides a clean evolution path from current architectures while bringing the control and automation required by AI.

EVPN complements segment routing by providing flexible, standards-based virtual networking for data centers and wide-area networks. It enables consistent policy and seamless workload mobility across locations, which is a necessity for agentic AI that often operates across multi-cloud and hybrid environments. Together, segment routing and EVPN create a converged network foundation that supports both traditional and AI workloads.

FlexAlgo for intelligent path selection

Adding segment routing and EVPN is not enough by itself. Networks also need to calculate the optimal path for different traffic types automatically. That is what FlexAlgo, short for flexible algorithm, enables. With FlexAlgo, operators can define separate algorithms for latency, bandwidth, resiliency, or even data sovereignty and then instruct the network to route different classes of traffic accordingly.

A network operator in healthcare might use one FlexAlgo instance to route medical imaging traffic on a guaranteed-bandwidth path, and another to route patient data along a path that remains within national boundaries. A financial services provider could use a low-latency algorithm for trading orders and a high-resiliency path for settlement traffic. These algorithms replace manual tunnel engineering and static policies with automatic, network-wide path computation. The benefits resemble those once pursued with RSVP-TE, but without the accompanying complexity, per-tunnel state and management burden.

In an AI-heavy environment, the number of traffic classes and SLAs continues to grow. FlexAlgo makes it possible to enforce a wide range of quality objectives on a single physical network. It also enables operators to quickly adjust performance parameters as changing business conditions require, a key attribute for supporting dynamic agentic workflows.

Security and policy in an automated network

Network automation does not reduce the importance of security; it makes security more configurable and more consistent. MACsec, for example, is now being integrated into modern IP networks to provide authenticated, encrypted communications at layer 2. This protects data moving between routers, switches and other network equipment from interception or tampering. For any AI deployment handling personal data or proprietary algorithms, such protection is essential.

Modern networking also allows security policies to be tied to application traffic automatically. As an AI agent connects to cloud services and on-premises resources, the network can classify the traffic, place it on the appropriate path and apply the correct set of access controls and inspection policies. This ensures that compliance is not dependent on individual administrators or on static rules that become obsolete in a rapidly changing application environment. It is particularly important in highly regulated industries, where data sovereignty and residency requirements require that certain traffic never crosses certain borders.

Real-world adoption in regulated industries

Forward-looking network teams are already incorporating these capabilities into their digital transformation strategies. In healthcare, for example, hospitals and research centers are under pressure to use AI for diagnostics, patient monitoring and operational efficiency. Their networks must carry imaging files, electronic health records, real-time telemetry and agent-driven workflows while remaining HIPAA-compliant and secure. Segment routing, FlexAlgo and MACsec together allow them to treat each traffic class based on its own requirements, without maintaining multiple parallel networks.

In finance, AI is increasingly used for fraud detection, customer service, risk analysis and algorithmic trading. These use cases demand low latency and precise traffic controls. A modern IP network with segment routing and FlexAlgo can meet these requirements while still supporting the many conventional applications that a bank runs. It can even facilitate new service offerings for providers. Depending on operational preferences, enterprises may build and manage such networks themselves over leased optical circuits, or they may buy them as a fully managed network service. The second model is particularly attractive as service providers look for new differentiators, creating an ecosystem of intelligent, policy-aware network services.

The strategic imperative of network evolution

Every enterprise needs to decide how quickly it will evolve its network. Agentic AI is not a temporary trend; it is becoming the primary interface for automation and data-driven decisions. At the same time, networks are not the seemingly neutral infrastructure they once were. They determine which applications can scale, which innovations can be delivered reliably and which markets a business can serve.

Modernizing IP networking is therefore a strategic decision, not merely a technology refresh. Organizations that recognize the critical role of connectivity will build the runway for AI innovation. They will be able to introduce new services in days, enforce service levels programmatically and establish trust with customers and regulators through built-in security and compliance.

The winners in the agentic AI era will be those that have cleared the network path before the demand arrives. The integration of real-time telemetry, segment routing, EVPN, FlexAlgo and MACsec security creates a networking foundation that can support AI agents wherever they operate, while preserving the stability required for traditional services. Those that take a proactive approach will be able to navigate the next wave of digital innovation instead of being constrained by the limitations of legacy infrastructure.

Source: Network World News

Arista's CEO Jayshree Ullal highlighted the company's rapid ascent, noting that just five years ago Arista generated roughly $2.9 billion in revenue for the entire 2021 fiscal year. "Customers see networking as the central nervous system for infrastructure from the client to campus to data and AI centers," Ullal said during the earnings call. The company also reported that its AI fabrics momentum with Etherlink switches now exceeds 100 cumulative customers, a sharp rise from the initial four or five customers Ullal spoke of in 2024.

Ullal said the company is seeing growth in virtually every sector, from back-end AI fabrics to the core data-center front end, and extending into adjacent campus and routing businesses. The scale-across switching and routing market is forecast to reach between $15 billion and $20 billion by 2030, positioning Arista well for continued expansion, she added.

Financial performance and aggressive growth forecast

The $3.036 billion quarterly figure represents more than just a milestone; it signals a fundamental shift in how enterprises and cloud providers are investing in networking. According to Arista's management, the company is now projecting 40% annual growth, which equates to an incremental $2.1 billion over its 2025 Analyst Day goal of $10.5 billion. That projection also marks an increase of $1.1 billion over the $11.5 billion target the company laid out in May 2026.

"We are now projecting 40% annual growth, and the market opportunity is broadening," Ullal said. The company's confidence comes from the sustained buildout of AI infrastructure, which requires high-bandwidth, low-latency networks capable of supporting massive distributed training and inference workloads. Unlike traditional data center traffic, AI workloads generate east-west traffic patterns that demand highly resilient, lossless fabrics.

Arista's financial results also reflect its ability to capture share in a market increasingly dominated by a few large players. The company has long competed against Cisco, Juniper, and Huawei, but its early focus on software-driven networking and telemetry has given it an edge in AI-era deployments. Analysts note that Arista's EOS operating system, which runs consistently across switching and routing platforms, reduces operational complexity and appeals to organizations running large-scale AI clusters.

AI networking momentum with Etherlink and beyond

Arista's Etherlink switch portfolio has become a cornerstone of its AI strategy. The company said it now has more than 100 cumulative customers using Etherlink platforms for AI fabric deployments. These customers range from large cloud service providers to enterprises building private AI infrastructure. "Our AI fabrics momentum with Etherlink switches now exceeds 100 cumulative customers from the initial four to five customers I spoke of in 2024," Ullal noted.

The growth is not confined to a single vertical. Arista is seeing adoption across financial services, healthcare, higher education, and the public sector, as organizations seek to build high-performance networks for AI model training, inference, and data processing. The shift toward AI-ready networking has also driven demand for higher-speed Ethernet, including 400G and 800G, as well as for technologies that improve utilization of expensive GPUs and XPUs.

Ullal emphasized that networking is now viewed as a strategic enabler of AI success rather than a commodity layer. "AI workloads are creating unprecedented pressure on network infrastructure," she said. "Arista's ability to deliver scalable, programmable, and observable networks is why customers are choosing us for their most demanding environments."

Supply chain pressures beginning to ease

One of the key questions hanging over Arista's recent quarters has been supply chain tightness. Last quarter, the company warned that shortages of memory, chips, and wafers were driving up costs and constraining its ability to meet customer demand. This quarter, management reported measurable improvements, thanks to a series of strategic actions taken over the past six months.

"Arista has spent the last six months improving our supply chain to meet growing product demand, and we're seeing significant improvements," said Todd Nightingale, president and COO of Arista. "We've secured multiyear agreements with leading vendors of strategic components, qualified new suppliers in key areas to limit risk, and built out supply chains for next-gen AI technologies."

Nightingale elaborated that relationships with strategic silicon vendors remain strong, with close collaboration on both supply chain and technical engagements. Memory supply for 2026 has been secured, and the company now has extended visibility well into 2027 across DDR4, DDR5, and NAND memory. "We've also increased our resiliency through optionality and expanded vendor qualification," he added. For printed circuit boards (PCBs) and optics, Arista can now build capacity within a 12-month window and has strengthened commitments from key suppliers.

The company has also improved lead times and inventory management for thousands of component SKUs. "We're improving sub-component pipelining and multi-sourcing, and providing increased flexibility with reduced inventory risk," Nightingale said. In addition, Arista has established a liquid-cooling supply chain designed to support the next generation of AI infrastructure. This includes cold plate, quick disconnect, and tubing vendors with capacity agreements for cutting-edge AI technology.

Ullal, however, injected a note of caution. "I don't want you to believe that suddenly we waved a magic wand and all our problems have gone away," she said. "The industry is going to have a two-year problem with memory and other silicon availability challenges, and I don't think we get out of it as an industry until 2028. But Arista is taking individually and specifically steps in the first half of this year that we believe will have results in the back half of this year."

EOS innovations: SSU, MRC, and SRv6

On the technology front, Arista's president and CTO Kenneth Duda highlighted three innovations within the company's EOS operating system that are helping differentiate its platforms in AI networks: Smart System Upgrade (SSU), Multipath Reliable Connection (MRC), and Segment Routing (SRv6).

SSU in an automated software upgrade feature designed to eliminate network disruption during OS updates. "Frequent upgrades are a hard reality today, especially as AI both uncovers security vulnerabilities and creates tools to exploit them," Duda said. "While many competing systems require a full reboot to address these issues, Arista EOS handles these upgrades seamlessly." This capability is particularly valuable in large AI clusters where downtime can stall expensive training runs and delay product development.

MRC addresses the problem of fabric cache collisions in AI networks. In first-generation AI networks, every packet on an XPU-to-XPU flow typically takes the same path. "That means if two flows hash to the same path, they both run at half speed," Duda explained. "MRC enables senders to spray a single flow across many paths through the fabric, where receivers reassemble any data that arrives out of order, eliminating the performance hit from fabric cache collisions." This capability is critical to maximizing the utilization of expensive accelerators, which often sit idle while waiting for data.

SRV6, a form of segment routing over IPv6, provides the control plane intelligence needed to direct traffic dynamically across the fabric. "It's not new, but using it to load balance an AI fabric—that's the game changer," Duda said. "The sender tags each packet with a stack of SRV6 segment IDs dictating the exact path the packet will take. The system then uses real-time congestion signaling to dynamically shift packets away from hotspots." This approach allows Arista to offer both high performance and operational simplicity.

Because Arista EOS is a single unified operating system running across the entire portfolio, the company can support SRv6 intelligence from the scale-out fabric to long-distance scale-across routing. "It gives our customers the combination of high-quality top performance and operational simplicity that Arista is known for," Duda added.

Optics and copper: a balanced approach

Optical connection technology continues to move into the AI networking environment, but Arista believes pluggable optics and copper will remain the dominant connectivity options through 2028-2029. Ullal expressed the company's philosophy succinctly: "Copper if you can, optics if you must."

Inside a rack, where distances are only two or three meters, copper cabling remains cost-effective and energy-efficient. For longer distances, pluggable optics are necessary. "I think you're going to see a lot of copper at those short distances, well within a rack," Ullal said. The company is less enthusiastic about proprietary implementations of co-packaged optics (CPO), where optical engines are integrated directly into the switch package. "Arista is not a fan of five different proprietary implementations," she stated.

Instead, Arista has been developing an open approach to co-packaged optics. "We don't think open CPO is going to happen overnight, but the idea here is to use socketed optical engines, pigtail fibers, and allow these modules to be fully pretested," Ullal explained. "And whether they're soldered on the board or nearby, the idea is to have a truly open interface that can operate with multiple vendors and multiple switch configurations."

Arista's support for open, socketed optical engines rather than multiple proprietary solutions is intended to give customers greater flexibility and reduce vendor lock-in. CPO and near-packaged optics (NPO) are expected to enter trials in 2027, but Ullal indicated they will remain a small portion of the market near term. In the meantime, Arista recently unveiled extended pluggable optics (XPO), a form factor designed specifically for high-speed optics. The company assembled more than 100 optics module suppliers as part of a multi-source agreement to build and support XPO, ensuring a competitive ecosystem and supply resilience.

Industry outlook and the road ahead

The AI networking boom shows no signs of slowing, and Arista is positioning itself to capitalize on the multiyear growth cycle. The company's raised guidance reflects confidence that its Ethernet-based AI fabrics will continue to displace older InfiniBand deployments, which have historically been favored for high-performance computing but lack the openness and scalability of Ethernet.

Ullal also pointed to the broader expansion of networking use cases, from campus and branch to data center and AI. "The scale-across switching and routing market is forecast to hit between $15 billion and $20 billion by 2030," she said. This includes not only high-speed data center switches but also the routers and aggregation devices that connect distributed AI infrastructure across multiple sites.

As AI models grow larger and training data sets expand, networks must evolve to handle unprecedented bandwidth demands. Arista's investments in supply chain resilience, software innovation, and open optical ecosystems are aimed squarely at that challenge. While supply chain constraints won't disappear overnight, the company's proactive steps and long-term vendor agreements provide a measure of stability that many competitors lack.

The road to 2028 will likely bring continued volatility in component availability, but Arista's record quarterly revenue and robust pipeline suggest that the company is managing the current environment effectively while positioning itself for the next wave of AI-driven networking growth.

Source: Network World News

The July 30 hearing, titled “Intelligent Networks: Powering Artificial Intelligence and Transforming Communications,” brought together a diverse group of witnesses, including representatives from U.S. Telecom, Vanderbilt University, and the Nebraska Public Service Commission. Lawmakers and industry experts alike explored how the rapid adoption of AI has stressed legacy network models and what steps are needed to ensure robust, secure connectivity in an AI-driven era.

In her opening statement, U.S. Senator Deb Fischer (R-Neb), chair of the subcommittee, set the stage by emphasizing the scale of the challenge. “We will explore how widespread AI use has forced networks to evolve, requiring more capacity and more complex designs so that AI can run efficiently on those networks,” Fischer said. She also noted that private companies have already invested hundreds of billions of dollars in network deployment, with additional federal broadband programs contributing billions more for targeted deployment and maintenance.

AI changes traffic patterns and network behavior

Everson’s testimony focused on a key insight: AI is not simply increasing the volume of traffic; it is changing the fundamental behavior of networks. “AI is changing not only the volume of network traffic, but the behavior,” he said. “Cisco measured a fourfold increase in AI inference traffic over eight months. Networks have traditionally been optimized for content flowing downstream. AI is far more two-way and uplink-intensive: prompts, context, sensor data, and agent activity all travel back toward AI models, and the resulting connections are active longer than conventional web transactions.”

This shift is particularly pronounced with the rise of AI agents—autonomous software that performs tasks without constant human supervision. Everson noted that in Cisco’s testing, an AI agent generated 450 percent more traffic than a person performing the same task, with roughly 70 percent of that additional traffic attributable to inference. These agents operate at software speed, creating bursts of traffic that traditional networks are not equipped to handle with conventional caching and downstream optimization techniques.

Campus and branch networks are already feeling the strain. Everson cited customer reports showing a 34 percent increase in traffic tied to AI workloads over the past year, with expectations of a 96 percent surge in the coming year. “Half of enterprise customers report that AI demand is concentrated on their Wi-Fi networks,” he said, adding that 73 percent of organizations already face or expect to face campus and branch capacity limitations within the next two years.

This is driven by significant increases in east-west traffic, latency-sensitive traffic, and continuous automated AI traffic. While much of the initial AI boom has relied on foundation models running in centralized data centers, Everson observed a growing trend toward distributed deployments. “We are seeing enterprises deploy more small language models, open-source models, and specialized models—such as vision and voice models—which can be distributed throughout the network,” he explained. This decentralization highlights the value of the Federal Communications Commission’s 2020 decision to authorize the full 6 GHz band for unlicensed Wi-Fi use, a move that Everson credited as forward-thinking.

Key areas impacted by AI infrastructure shift

Everson’s prepared remarks laid out several key areas where AI is forcing network operators to rethink their strategies, spanning infrastructure, technical performance, cost, and data governance.

• Infrastructure: AI is accelerating the move toward edge computing. Everson argued that service providers must plan for “AI-native” traffic profiles, considering technical demands, cost implications, and growing concerns about data sovereignty and security. Rather than backhauling everything to centralized clouds, operators should push compute capacity toward the network edge—at cell sites, for example—to enable faster processing and reduce bottlenecks.
• Technical performance: Emerging physical AI applications, including robotics, autonomous vehicles, and industrial automation, may require sub-millisecond decision-making. Everson gave a vivid example: if an autonomous robot sends data to a central cloud and must wait for a round-trip response, the latency could be too high for safe, real-time operation. Localized inference and edge processing become essential to meet such demanding performance thresholds.
• Cost: AI operations generate enormous amounts of data. High-definition video analytics for public safety alone can create terabytes of data each day. Transporting that data to a centralized processing hub is not only expensive but also risks overwhelming network backhaul links. Edge processing and distributed AI architectures offer a more economical alternative, reducing congestion and enabling scalable growth.
• Data sovereignty and security: Enterprises and government agencies exhibit growing anxiety about where sensitive data resides. Many customers have security or regulatory reservations about sending proprietary or classified information across public networks to third-party cloud providers. Edge computing can help keep data within local jurisdiction, alleviating compliance burdens and strengthening security posture.

AI as a tool for network resilience

While AI workloads pose several challenges, Everson also emphasized the tremendous upside of using AI to enhance network operations. “There is a tremendous opportunity to leverage AI to deliver new applications and better performance, infuse security into the fabric of the network, and manage the increased complexity,” he said. In particular, he highlighted the role of agentic AI in changing both network traffic patterns and the administrative tools available to operators.

One such advancement is AgenticOps, which enables networks to act as self-healing systems. Everson described practical applications where Cisco’s AI-native tools automatically reroute traffic, adjust capacity, or reconfigure network nodes when performance degradation or impending hardware failure is detected. “This dramatically increases uptime and reliability for mission-critical services,” he asserted.

In addition to operational improvements, AgenticOps can help bridge the network industry’s talent gap—a persistent issue cited by many IT leaders. By automating repetitive, low-value tasks such as ticket resolution, configuration updates, and routine maintenance, network operators can lower the bar for entry-level professionals. “These tools also help close the workforce talent gap by lowering the barrier to entry and allowing more junior analysts to ramp up quickly,” Everson said. Freed from mundane chores, experienced engineers can concentrate on architectural strategy, while cybersecurity analysts can spend more time on threat hunting and detection engineering.

AI-native platforms and integrated sensing

The evolution of networks is not limited to incremental upgrades; Everson outlined a more fundamental transformation. Networks are moving from simple data pipes to AI-native platforms that function as intelligent connectivity fabrics. As compute moves toward the edge, operators can host applications directly on network infrastructure, unlocking new capabilities.

A notable example discussed by Everson is Integrated Sensing and Communication (ISAC). This technology merges wireless communications with radio-frequency sensing to detect the position and path of objects using reflections of radio waves. ISAC offers significant advantages over traditional optical sensors: it can penetrate smoke, operate in low-light conditions, and see around obstacles where conventional video analytics would fail. “This technology has been prototyped and demonstrated already, and it holds great promise for autonomous systems and robotics, AI-driven smart facilities, and public safety,” Everson noted.

The implications of ISAC and other edge-based innovations extend far beyond commercial efficiency. They are foundational capabilities for future smart cities, intelligent transportation, and next-generation public safety solutions. As these systems mature, they will require robust, low-latency, and secure networks that can support both communications and sensing functions simultaneously.

Policy recommendations for the committee

Everson concluded his remarks by offering three concrete suggestions for lawmakers to consider as they chart the future of national technology policy.

• Accelerate the U.S. AI-native stack. Cisco is investing heavily in AI-native networking capabilities, from 5G-Advanced innovations today to building blocks for 6G. A prominent example is AI-WIN, a collaborative venture involving Cisco, NVIDIA, MITRE, Orion Development Company, Booz Allen, and T-Mobile. The project marries AI, compute, and wireless technology to create a secure, American-led path from 5G-Advanced to AI-native 6G. Everson encouraged Congress to “lean in” on strategic areas such as compute, core networking, and applications where the United States has a unique edge.
• Modernize permitting and infrastructure rules. As computing becomes more distributed, outdated permitting processes could slow network deployment. Everson urged the committee to address these hurdles while also reconsidering the Universal Service Fund to reflect the evolving costs of AI-ready networks. His concern is that rural and disadvantaged communities might otherwise miss out on the benefits of advanced AI connectivity.
• Maintain a balanced spectrum policy. Everson emphasized the need for both licensed and unlicensed spectrum. The 800 megahertz of licensed spectrum recently made available by Congress is crucial for high-capacity, uplink-intensive applications, while the FCC’s authorization of the 6 GHz band for unlicensed use is equally important to meet enterprise demand. He insisted that “a dependable pipeline of both is foundational to American leadership” and thanked the committee for its ongoing efforts to rebuild that pipeline.

Everson’s testimony underscores a pivotal moment for the networking industry. As AI continues to infuse every layer of the technology stack, the systems that carry data must become smarter, faster, and more flexible than ever before. The debate in Washington—and the decisions that follow—will help determine whether the nation can build the infrastructure needed to sustain AI leadership, strengthen economic competitiveness, and ensure equitable access to the digital future.

Source: Network World News

Groundcover builds observability technology on the open-source eBPF and OpenTelemetry technologies. Founded in 2021, the company raised $35 million in a Series B round in April 2025 and has spent the time since extending that foundation to cover AI agents and the tools those agents call in production. The new funding round is led by One Peak, with participation from Morgan Stanley Expansion Capital, Zeev Ventures, Angular Ventures, Heavybit, and Jibe. The investment brings Groundcover's total funding to $160 million.

“I think what is happening to observability right now is fascinating,” Groundcover CEO and co-founder Shahar Azulay said. He sees the industry moving beyond traditional application performance monitoring toward a broader discipline that must account for AI-driven software development and operations.

What eBPF does and why it matters more now

eBPF, short for extended Berkeley Packet Filter, is a Linux kernel technology that lets code run safely inside the kernel without a custom kernel module. It has long been used for network monitoring. Groundcover uses eBPF to watch application and infrastructure activity without requiring a developer to instrument each service by hand. That approach removes a step most observability vendors still require. “You didn’t have to have the developer instrument an SDK, change their code base, and so on,” Azulay explained.

The same property is becoming useful for a different reason now. Engineering teams are adopting new AI tools fast enough that they lose track of what is actually running in their own environment, Azulay said. He compared the gap to the visibility problems teams dealt with roughly a decade ago, before observability tooling matured. eBPF operates below the application layer rather than depending on code a developer wrote, so Groundcover can still see workflows nobody thought to instrument.

“eBPF is kind of that security net of even if you didn’t instrument, even if you’re not in full control, you’re gonna know which agentic workflows are running in production, which models are using, which vendors they’re using, and so on,” Azulay said. This visibility is essential in environments where AI agents can be introduced by different teams, often with little coordination.

The use of eBPF also gives Groundcover a lightweight footprint. Because it does not require agents or SDKs, the technology is less intrusive and can monitor high-throughput production systems with lower overhead. That is a significant advantage for enterprises moving to cloud-native architectures where traditional monitoring tools can become heavy and costly.

How agentic workflows are breaking distributed tracing

Distributed tracing follows a request as it moves across services so engineers can see where time is spent and where something broke. It has always relied on a predictable number of hops, the kind of path an engineer could trace by hand, such as a cache calling a database. In conventional microservices, a trace is essentially a finite tree with known service boundaries. Engineers can use that structure to identify bottlenecks, error propagation, and latency cascades.

Azulay said that assumption breaks down once agents enter the picture, since a single agent session can generate a large number of tool calls and internal model calls with no fixed pattern. “With LLMs and agentic workflows, this is becoming very complicated,” Azulay said. An agent may call an external API, then decide to query a vector database, then invoke a code generation tool, then synthesize results, and then start a new chain of calls based on the output. The trace is no longer a predictable series of service-to-service requests but a branching, adaptive graph that reflects the model’s decisions at each step.

This complexity makes it difficult to apply traditional root-cause analysis. When something goes wrong, it may not be a single service failing but a sequence of AI decisions that led to a poor outcome. Even if the individual calls succeed, the user-visible result can be wrong, misleading, or harmful. Observability systems need to capture not just the calls, but the reasoning context around them.

Teams now also track token usage and hallucination rates alongside latency and error rate, Azulay said. Token consumption has become a financial and operational concern because AI systems are billed per token, so runaway usage can lead to unexpectedly high costs. Hallucination rates measure how often the model produces factually incorrect or unsupported content. These are new metrics that traditional observability tools were not designed to handle.

Traces can contain a customer’s actual prompt instead of only structured request data, which raises privacy questions. If a user submits sensitive text to a prompt and that prompt becomes part of a trace, the observability system must protect that data at least as carefully as the application itself. Azulay does not consider the result a variant of application performance monitoring. “It’s not going to be the same product,” he said. “AI observability is not exactly APM.”

The distinction matters for vendors because AI observability demands new data models, new storage systems, and new privacy controls. A trace is no longer just a sequence of spans; it can include model outputs, context windows, tool definitions, and the state of an agent’s memory. All of that needs to be captured and correlated to provide useful insights.

Azulay tied that shift back to Groundcover’s own architecture. Because the platform stores telemetry inside the customer’s own cloud environment rather than a shared vendor backend, he said it is built to hold the larger, more sensitive telemetry volumes agentic workloads produce without shipping that data to a third party. That approach can reassure enterprises with strict data residency requirements and compliance obligations. It also allows them to retain full control over who has access to the data and for how long.

“I think people are expected to save more telemetry, and save more telemetry more privately,” Azulay said. The shift to tenant-controlled storage is a notable departure from the centralized SaaS model adopted by many observability platforms. But Groundcover believes it is the right architecture for AI-related workloads because the data is not only bigger, it is also semantically richer. In many cases, the telemetry itself may contain intellectual property, trade secrets, or personally identifiable information.

Agent Mode and the rise of MCP

Groundcover isn’t just optimizing its platform for the needs of modern agentic AI activity. The company is also using AI to improve user experience. Agent Mode is Groundcover’s built-in AI assistant for engineers, letting them ask questions about their systems, build dashboards, and troubleshoot problems in logs and traces without writing queries by hand. Instead of querying with PromQL or other query languages, a developer can type a question in natural language and get an answer with supporting data.

Groundcover has also built a Model Context Protocol (MCP) integration that connects Agent Mode to coding agents and workflow tools including Linear, letting engineers and AI systems pass context back and forth during an incident. MCP is an open standard that enables AI models to interact with external tools in a structured way. It has been adopted by a wide range of AI-assisted development platforms, making it easier for observability tools to plug into modern developer workflows.

Azulay said adoption of MCP integration has moved faster than the company expected. Customers use the integration differently depending on how far along they are in adopting AI tools, according to Azulay. Some ask questions through it instead of opening the Groundcover dashboard, while others use it to write a fix directly. In the first case, the MCP endpoint serves as a bridge to the observability platform’s natural-language capabilities. In the second case, the AI coding agent uses MCP to gather context during an incident and then proposes code changes that a human can review.

He framed the pattern as an industry trend rather than a product win specific to Groundcover. “People are basically starting to build their autonomous software development structure,” he said. This structure involves AI agents that not only write code, but also test, debug, and reason about production systems. For that to work, the agents need a rich and reliable source of telemetry data that describes what is happening across the entire stack.

Groundcover’s focus on AI infrastructure suggests a broader transformation in the observability market. It is no longer enough to monitor metrics, logs, and traces in isolation. Modern systems require an integrated understanding of how AI models interact with standard software components. The company is positioning itself to be the visibility layer for that new reality.

Groundcover at a glance

• Founded: 2021
• Total funding: $160 million
• Latest round: $100 million Series C, led by One Peak
• Other investors: Morgan Stanley Expansion Capital, Zeev Ventures, Angular Ventures, Heavybit, Jibe
• Headquarters: Tel Aviv, Israel
• CEO: Shahar Azulay
• What they do: Observability technology built on eBPF and OpenTelemetry

With the new capital, Groundcover expects to accelerate its product roadmap and expand its go-to-market efforts. The company is betting that the observability market will revolve around AI infrastructure for the next several years. As more organizations adopt agentic workflows and large language models in production, they will need to understand what these systems are doing, how they perform, and whether they are operating within safety and privacy guardrails.

The emergence of AI agents has also brought new roles and responsibilities for platform engineering teams. In the past, observability was primarily a concern for SREs and DevOps engineers. Now, it is becoming relevant to AI engineers, ML platform teams, and even legal and compliance groups that need to audit model behavior. Observability platforms must serve all of these audiences with different views and controls. Groundcover’s approach is to provide a unified data layer that can answer both technical and behavioral questions about AI systems.

Industry observers have noted that the observability tooling market has grown crowded over the past few years. Established players are adding AI-driven features, while startups try to differentiate on cost, ease of use, or architectural innovation. Groundcover believes its eBPF foundation is a meaningful differentiator because it addresses the fundamental challenge of collecting reliable data without interfering with modern development velocity. The company’s use of OpenTelemetry is another important factor, as it ensures interoperability with the broader ecosystem and avoids vendor lock-in.

Groundcover’s rise also reflects a broader trend in which infrastructure software is being rebuilt around AI. Cloud providers have been offering instance types with increasingly powerful GPUs, but software that helps developers manage and observe those systems is only starting to catch up. Observability is one of the most logical places for AI integration because the volume of data is high, the queries are often complex, and the need for real-time insights is urgent. AI can help sift through logs, identify anomalies, and even suggest root causes that a human might have overlooked.

The challenge is that AI itself creates new kinds of problematic behaviors. Hallucinations, prompt injection attacks, and non-deterministic outputs are not well-handled by standard monitoring tools. A prompt injection might cause an agent to perform unauthorized actions, and without proper tracing that behavior could go unnoticed. Observability platforms need to detect that kind of subtle failure, which requires inspecting the inputs and outputs of AI models. That data is far more sensitive and unstructured than typical application logs, making the storage and processing challenge even more complex.

Azulay’s philosophy is that observability must evolve from recording what happened to explaining why it happened. Traditional monitoring tells you that a service is down; observability tells you why. The coming generation of AI observability will tell you why an AI agent made the decisions it did, what context it drew from, and which influences led to a specific outcome. That level of understanding is necessary for trusting AI with critical business functions.

The investment community has taken notice. Large funding rounds in the observability sector have become less common as the market matures, but Groundcover’s $100 million round suggests that investors see opportunity at the intersection of AI and infrastructure. The company’s customer base has been growing, and its platform has been adopted by organizations that need visibility into both conventional services and AI workloads. Groundcover says it will use the funding to hire more engineers, deepen its AI capabilities, and strengthen its presence in global markets.

The company is also focused on expanding its partner ecosystem. Many enterprises use multiple cloud providers and a wide range of software as a service tools. Groundcover works within that diversity by supporting open standards and providing integrations with popular developer platforms. The MCP integration is just one example of the company’s push to make observability data available where engineers already work, whether that is a chat interface, a code editor, or an incident management platform.

As AI continues to permeate every layer of the technology stack, the boundary between applications and infrastructure is blurring. Soon, most business software will include machine learning components in some form. That increases the need for developers to understand not only the code they write but also the models their code calls and the data those models process. Observability is the natural foundation for that understanding. Groundcover is seeking to lead in that new era, armed with a technology approach designed for complexity, privacy, and scale.

Source: Network World News

The backlog was disclosed alongside quarterly results that showed nearly every part of Dell’s infrastructure business is now being reshaped by AI. Revenue climbed at a double-digit rate, infrastructure sales hit a record, and the order pipeline continued to expand even as customers waited longer for components. Dell’s chief operating officer, Jeff Clarke, said supply constraints begin with servers and storage and stretch across the technology stack to “just about every product going through a leading node.” He stressed that the company is doing all it can to secure more inventory and acknowledged that, in the current environment, that is a difficult task.

Quarterly numbers reveal accelerating infrastructure demand

Dell’s fiscal quarter ended July 31 with revenue of $47 billion, a 58% increase over the same period a year earlier. The Infrastructure Solutions Group, which houses the company’s servers, storage, and networking products, generated $31.8 billion in revenue, an 89% jump and a record for the division. Server-related earnings were particularly strong, rising 122% year over year, and Dell said demand for CPU-based servers is “exceptionally strong” as those systems increasingly support agentic AI workflows.

The order book shows no sign of cooling. Dell booked almost $61 billion in AI server orders in the three months through July 31, and the company recorded more than $130 billion in AI server orders over the prior 12 months. According to Clarke, Dell converted $131.7 billion of customer demand into purchase orders during that period. The sources of that demand are widening, he said, with enterprise customers, neocloud providers, and sovereign cloud operators all competing for capacity.

What agentic AI means for the data center

Agentic AI refers to systems that can reason through complex tasks, plan next steps, call external tools, and operate with less direct supervision than conventional chatbots or copilots. Every one of those steps often requires another inference pass, more data retrieval, and additional model context. That makes agentic workloads far more demanding than earlier AI applications, and it is forcing infrastructure teams to think differently about capacity planning.

Dell’s long-range projections paint a dramatic picture. Clarke said the number of tokens used could reach 3,600 quadrillion by 2030, an 87-fold increase from current levels. Training demand is expected to rise to 850 zettaflops over that period, a fivefold increase. The more consequential shift may come from inference, which Clarke described as “pure demand” in the industry. He predicted that enterprise agentic AI will become the single largest data center workload by 2028 and could account for 75% of all data center demand by 2030.

Traditional servers are back in the growth story

One of the most striking findings in Dell’s report is how much of the recent expansion is tied to traditional CPU-based servers. It has been easy to assume that AI infrastructure is mostly about GPUs and specialized accelerators, but Dell says customers also require “meaningful CPU compute capacity” to support their AI and agentic workflows. In just the last two fiscal quarters, Dell generated nearly as much revenue from traditional servers and networking as it had in any previous full year in company history, according to Clarke.

A large part of that growth is coming from existing customers accelerating their investments in conventional IT environments. Enterprises are refreshing servers to modernize, improve performance, and become more resilient. Security requirements are also adding urgency. Dell expects these refreshes to be both significant and durable, especially as businesses prepare their data centers for workloads that use more cores, more dynamic random-access memory, and more storage capacity.

Dell also emphasized that AI infrastructure is not simply a matter of assembling components and shipping a rack. Modern AI deployments require disaggregated architectures that keep data accessible and in motion across compute, storage, and networking. Clarke said some customer engagements require more than 50 unique server designs because enterprises must optimize for workload performance, power efficiency, cooling, and the physical data center environment. That engineering burden is becoming one more reason why the infrastructure market cannot respond quickly to sudden demand spikes.

The supply chain remains the limiting factor

Even with rising demand, the biggest obstacle for Dell is not customer interest. It is the global component supply chain. Clarke described the pressure in unusually direct terms, saying the constraints begin with DRAM and memory, then extend to NAND flash and a long list of other parts. He summed up the moment as “DRAM, DRAM, DRAM, followed by NAND, NAND, NAND.”

The shortage is not limited to the highest-profile AI components. Dell said there are “spotty” shortages in CPUs and disk drives, and constraints reach down through microcontrollers, drives, transistors, and the broader electronics supply base. The memory shortage is especially painful because new AI servers require more DRAM per system than typical enterprise servers. High-bandwidth memory used in accelerators has also absorbed a large share of memory industry output, making the squeeze even more visible in other parts of the market.

Those shortages limit how many complete AI systems Dell can build in any given quarter. Clarke said large enterprises and multinational corporations around the world would prefer to receive products immediately if supply allowed. “We are supply constrained in the sense of what we can build in any given quarter,” he explained. The company is working through lead times and trying to optimize the configurations it can produce with available components.

Storage is becoming another bottleneck

Agentic AI depends on continuous access to large data sets, which makes storage a strategic concern rather than an afterthought. Dell reported strong growth across its PowerFlex, PowerStore, PowerProtect, and PowerVault product families. The company sees steady demand as organizations modernize storage environments to handle growing volumes of operational data and AI training data.

Clarke said demand remains broad based and that enterprises are increasingly focused on keeping data available and secure while also moving it efficiently across compute and networking resources. AI systems need to read and write data constantly during both training and inference, so storage latency and throughput can quickly become a performance ceiling. The record order backlog suggests that many enterprises have recognized this and are trying to build AI-ready storage before their next wave of model deployments begins.

How customers are responding to higher costs and longer lead times

The shortage-driven environment is also affecting pricing. Clarke noted that modernized configurations with higher core counts, more DRAM, and more storage cost more than they did in previous quarters. Some customers are deferring purchases because they cannot stretch existing budgets enough to cover the rise. Others are placing orders much further in advance just to guarantee access to constrained components.

Large, sophisticated customers are often the first to act, according to Clarke. He said some are collaborating with Dell on longer-range planning so they can map their infrastructure needs well into the future. Clarke described that behavior as “a new phenomenon” in the market and said Dell is working to help customers manage in an environment where demand is running well ahead of supply.

In the meantime, Dell continues to make operational adjustments. The company is optimizing configurations with the components it has and focusing on getting systems out the door as quickly as possible. Clarke said Dell has been able to increase shipments despite the constraints, but the effort requires constant attention to supply allocation, engineering resources, and customer priorities. The $95 billion backlog is likely to remain a central metric for the AI infrastructure market until manufacturing capacity catches up with the orders already on the table.

Source: Network World News

When asked directly whether James Gunn and Peter Safran ever told her they planned to take the character in a different direction, Gadot said that was not what she heard during their conversations. “They told me the opposite!” she said. “They told me, ‘We’re going to develop it with you. The third [movie]...’”

Gadot previously shared a similar account in 2023 interviews, saying she had met with the new heads of DC Studios and was reassured about her future in the role. She recalled being told, in what she described as a direct quote: “You’re in the best hands. We’re going to develop ‘Wonder Woman 3’ with you. [We] love you as Wonder Woman— you’ve got nothing to worry about.”

At the time, people with knowledge of the situation pushed back on that characterization. The official word from within the studio was that Wonder Woman 3 was not actively in development, that Gadot had not been promised another sequel, and that no definitive conversation had taken place about the character continuing into the new DC Universe.

The conflicting accounts have remained a source of confusion for fans ever since. Gadot’s latest remarks add another layer to that history because she still maintains that the new DC leadership initially signaled she would remain involved.

Gadot first entered the world of DC as Wonder Woman in Zack Snyder’s Batman v Superman: Dawn of Justice. She went on to lead two standalone films directed by Patty Jenkins: 2017’s Wonder Woman and 2020’s Wonder Woman 1984. Gadot also appeared alongside the rest of the Justice League in the 2017 ensemble film and made cameo appearances in other DC titles, including Shazam! Fury of the Gods.

Before Gunn and Safran arrived, a third Wonder Woman movie was in development with Jenkins returning as director and Gadot expected to play the title role again. That project had become one of the anchor movies of DC’s old shared universe, especially after the first film became a cultural milestone and a major box office success. The sequel was released during the pandemic and faced a more complicated reception, but there was still enough interest in the character that a threequel seemed like a natural next step.

By the time the new DC leadership was installed, the studio was in the middle of a broad reset. Many of the actors associated with the earlier era were being phased out, and Wonder Woman was not the only high-profile casualty. Gadot said she ultimately was not shocked by the outcome because she had watched how other exits were handled.

“I saw the way everybody else was handled,” she explained. “The Henry situation was not handled elegantly, to say the least. Same with Patty. It was not handled elegantly. We’re all adults and it’s all good. It’s not even about [James Gunn and Peter Safran]. They are there and they’re there to do what they believe in. It’s all good.”

Gadot’s mention of Henry Cavill points to one of the messiest parts of the DC transition. Cavill returned as Superman in the post-credits scene of Black Adam, and his return was announced publicly with a sense of excitement. Fans who had campaigned for Cavill to get another chance as Superman saw that as a major victory. But the new DC regime soon made it clear that Cavill was not part of its plans.

Instead, the studio moved forward with a fresh version of Superman played by David Corenswet. Cavill’s exit became a major talking point because it followed such a public return. Gadot’s comments suggest she saw that sequence of events as emblematic of a wider issue with how the original DC era was being wound down.

Patty Jenkins’ exit from Wonder Woman was also awkward. News broke in December 2022 that the third Wonder Woman film had fallen apart after studio leadership passed on Jenkins’ treatment for the sequel. The timing was notable because it happened just as Gunn and Safran were taking over DC Studios, which led many to assume the project was being cleared away for the new direction.

Jenkins pushed back on the idea that she had left willingly. She said at the time: “I never walked away. I was open to considering anything asked of me. It was my understanding there was nothing I could do to move anything forward at this time. DC is obviously buried in changes they are having to make, so I understand these decisions are difficult right now.”

Jenkins later said she believed the Wonder Woman story was over “for the time being, easily forever.” She expressed sympathy for the enormity of what Gunn and Safran were trying to build, but she also made it clear that the decision had not come from her side. “They aren’t interested in doing any ‘Wonder Woman’ for the time being,” Jenkins said at the time. “It’s not an easy task, with what’s going on with DC. James Gunn and Peter Safran have to follow their own heart into their own plans. I don’t know what they are planning on doing or why, so I have sympathy for what a big job it is and they have to follow their heart and do what they’ve got planned.”

Gadot’s latest interview suggests she agrees with the idea that Gunn and Safran had a different creative vision and were not simply trying to undo the work of the people who came before them. At the same time, she was honest about the fact that the studio’s communication could have been better. Her tone throughout the conversation was measured rather than bitter, even as she expressed frustration with the lack of clarity.

Asked about the possibility of returning to Wonder Woman in the future, Gadot did not seem especially eager to go back. She pointed to her personal life as the main reason. Gadot has four children, and she explained that taking on a major superhero movie would require a significant amount of time away from them. “For me to go to do another Wonder Woman, I am not sure I would at this point in my life with four children and different schools and all of that,” she said. “To check out from all of them for such a long period of time, it doesn’t work for my life anymore at this point.”

That answer is a notable shift from the years when Gadot seemed fully committed to the idea of continuing as Wonder Woman. The role made her one of the most recognizable stars in the superhero genre and gave her a franchise that few other actors in the DC universe had. Still, she appears to have accepted that this chapter is over and that her priorities have changed.

Her comments also capture some of the difficulty of reboot culture, where actors can be told one thing by one wave of executives and something very different by the next. The shift at DC Studios was always going to leave some actors feeling caught in the middle. Gadot has now made it clear that she felt that tension herself, even if she ultimately holds no hard feelings toward the people now in charge.

For fans of the older DC films, the idea of a Wonder Woman 3 with Gadot and Jenkins together will remain one of the great what-ifs of that era. The first film proved that a solo Wonder Woman movie could be both commercially successful and emotionally resonant. Gadot’s casting was widely praised, and Jenkins brought the legendary character to the screen in a way that felt distinctly different from the other

Source: Variety News