Chamath Palihapitiya, the founder of Social Capital and a prominent tech investor, has released an AI investing guide that maps out where he expects money to flow across the artificial intelligence market. In a series of posts on X, he told his followers that the fastest path to cash-on-cash returns lies in land, power, and shell (LPS), but he also made a bold case for why the real long-term winners will be harnesses and the applications built on top of them.
The guide comes at a time when the AI industry is experiencing explosive growth, with massive capital expenditures pouring into data centers, chips, and model development. Palihapitiya, known for his early investments in companies like Facebook and his outspoken views on tech markets, is now turning his attention to the infrastructure and software layers that will define AI's next phase.
The LPS Layer: Land, Power, and Shell
Palihapitiya categorized his AI investments into layers, and the first layer he addressed is LPS, which stands for land, power, and shell. This refers to the physical footprint of a data center before any chips are installed. In his post, he described this layer as "still the most obvious and fastest path to cash on cash returns." He added that "lots of value can be assembled and traded quickly at this layer. And as data centers get more pushback, energized land can explode in value. Very bullish here."
The emphasis on land and power reflects a growing bottleneck in the AI supply chain. As hyperscale cloud providers and AI startups race to build out computing capacity, they are encountering constraints in securing enough land with access to reliable electricity. Palihapitiya noted that he and his partner, Anita Verma-Lallian, have locked in close to 6 gigawatts of power running through 2029. This is a significant amount of energy, enough to power millions of homes, and it positions them to capitalize on the surging demand for data center capacity.
Palihapitiya had previously stated that zoning-approved land and silicon access hand their owners negotiating leverage over everyone downstream. In other words, whoever controls the physical and electrical resources of data centers holds substantial power over the rest of the AI value chain. This insight aligns with broader industry trends, where data center developers are increasingly competing for access to power grids and suitable sites.
The Harness: Software That Unlocks AI's Potential
Above the concrete and the power lines, Palihapitiya's pick is the harness. In a post made in July, he explained, "A modern harness + open model will crush your token consumption but keep your performance." A harness, according to a Hugging Face glossary, is the software wrapped around an AI model that decides what the model sees, which tools it can call, and when it stops. In essence, the harness acts as the controlling layer that directs the AI model's behavior and integration with enterprise systems.
Examples of harnesses include Anthropic's Claude Code, OpenAI's Codex, and Google's Antigravity. Claude Code is explicitly referred to as "the agentic harness around Claude" in its documentation. These tools are becoming critical as companies move from simple chatbot interactions to complex, agentic workflows that involve multiple steps, tool usage, and integration with proprietary data.
Palihapitiya argued that "the harness helps enterprises own their proprietary context (what Alex Karp calls their 'alpha')," which includes their data, workflows, business rules, and other unique assets. This is a key differentiator because while AI models are becoming commoditized, the context and control layers that make models useful for specific businesses remain proprietary and defensible.
The importance of the harness cannot be overstated. As AI models become cheaper and more interchangeable, the value shifts to the software that orchestrates them. A well-designed harness can significantly reduce token consumption, thereby cutting costs, while maintaining or even improving performance by optimizing how the model is prompted and which tools it accesses.
Applications: The Ultimate Destination for Value
Palihapitiya's thesis extends to applications, which he believes will be another long-term winner. He wrote, "Every company, with the right harness, can now imbue their alpha into the software that runs their company." This suggests that the ability to customize AI systems with proprietary knowledge will enable companies to build powerful applications that were previously impossible or prohibitively expensive.
Applications built on top of harnesses and models represent the final layer of value creation. They are where the rubber meets the road, delivering tangible outcomes for businesses and consumers. Palihapitiya's view is that the AI market will not be dominated solely by model makers or infrastructure providers, but by the companies that successfully leverage harnesses to build applications tailored to specific industries and use cases.
This perspective is gaining traction in the tech community. Xiaoyin Qu, the founder of Tycoon AI, expressed strong support for harnesses, stating that a harness "will create margin regardless of if the model gets commoditized," because the right one can unlock large, long-horizon jobs that are worth more than any single model output. Aaron Levie, Box's CEO, echoed this sentiment in response to a different post, saying that the harness is "going to become the most important variable" in the AI stack, sitting right next to raw model capability.
Industry Reactions and Context
Palihapitiya's guide has sparked discussion among industry leaders. Many have chimed in to support his views, especially his emphasis on harnesses. The consensus is that as AI models become more commoditized, the ability to create value will increasingly rely on the software that controls and applies those models to real-world problems.
This is a notable shift from earlier discussions that focused heavily on model quality and scale. While model capability remains important, the harness is emerging as a critical differentiator. It allows organizations to maintain control over their data and workflows, which is essential for compliance, security, and competitive advantage. Alex Karp, CEO of Palantir, has long been an advocate for this idea, referring to the unique data and processes that give companies their "alpha." Palihapitiya's embrace of this concept signals a broader recognition that AI's true value lies in its application to specific business contexts.
Palihapitiya's guide also arrives at a time of intense debate about AI spending. In mid-July, he made a post on X questioning whether current AI spending was paying off for anyone beyond the handful of firms already collecting the money. He pointed to buyers who can now spend $0.50 per million leading-edge tokens instead of $56 for the same volume, a dramatic price decline that reflects the rapid commoditization of AI models. The harness call appears to be his answer to this concern: if models get cheap and interchangeable, the money moves to whoever controls the data, workflows, and applications built on top.
Palihapitiya's background gives weight to his investment thesis. As a former senior executive at Facebook and the founder of Social Capital, he has a track record of identifying transformative trends early. His investments have spanned technology, healthcare, and finance, and he has been a vocal advocate for using capital to solve big problems. His focus on AI infrastructure and software layers is consistent with his broader investment philosophy of betting on fundamental shifts in the economy.
The AI market is currently in a phase of massive expansion. According to industry estimates, spending on data centers, power, and related infrastructure is expected to grow significantly in the coming years, driven by the demands of AI workloads. The competition for land and power is intensifying, with data center developers exploring new locations and negotiating with utilities for access to electricity. Palihapitiya's early positioning in this space, with 6 gigawatts of power secured, puts him at the forefront of this trend.
At the same time, the software layer is evolving rapidly. Harnesses are becoming more sophisticated, with capabilities that allow AI models to interact with corporate systems, access databases, and execute complex tasks. The rise of agentic AI, where models operate autonomously to achieve goals, is further boosting the importance of harnesses. Companies like Anthropic, OpenAI, and Google are investing heavily in these tools, recognizing that they are essential for making AI practical and usable in enterprise settings.
Palihapitiya's guide serves as a roadmap for investors and entrepreneurs looking to navigate the AI landscape. It highlights the opportunities at each layer of the stack, from physical infrastructure to software to applications. By emphasizing the importance of harnesses, he is pushing back against the notion that AI is solely about model size and compute power. Instead, he suggests that the winners will be those who build the connective tissue that makes AI work for specific businesses.
The reaction from industry insiders suggests that Palihapitiya is onto something. The consensus is that while models will continue to improve, they will become increasingly commoditized. The real value will accrue to those who can control the context and application of AI, which is precisely what harnesses and applications enable. This is a nuanced and forward-looking perspective that challenges conventional wisdom about where the money in AI will be made.
As the AI industry continues to mature, Palihapitiya's guide is likely to be seen as an important contribution to the conversation about AI investing. His calls on land, power, and shell, as well as harnesses and applications, provide a framework for understanding the various components of the AI ecosystem and how they will generate returns. For investors, the key takeaway is that the fastest cash may be in infrastructure, but the lasting value will be in the software that unlocks AI's potential for enterprises.
The evolution of AI is far from over, and Palihapitiya's insights offer a glimpse into where the market is heading. With billions of dollars flowing into AI ventures, having a clear strategy is essential. His guide, with its emphasis on layered investments and the critical role of harnesses, provides a valuable roadmap for those looking to participate in the AI revolution. While no one can predict the future with certainty, Palihapitiya's track record and the early support from industry leaders suggest that his perspective deserves careful consideration.
Source: MSN News