Observability spent most of the past decade as a post-production discipline, catching outages and cutting the time engineers need to find a root cause. That focus is shifting as agentic AI systems move into the software development lifecycle, pulling production context earlier into coding, testing, and deployment work. This shift is helping to fuel growing demand for observability vendor Groundcover, which this week announced a $100 million Series C round.

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