Chamath Palihapitiya backs harnesses and applications in AI investing guide

Investor Chamath Palihapitiya has published a map of where he expects money to move through the AI market on X, telling his followers that the fastest cash sits in power and data-center real estate.

However, he also stated that the durable margins will belong to “harnesses” and the applications built on top of them.

Where did Palihapitiya say the cash will land first in his guide?

The Social Capital founder tagged his post on X as his “AI investing guide” as of August 2026. 

He categorized areas for investments into layers, and the first layer he treated was what he calls LPS, short for land, power, and shell. It refers to the physical footprint of a data center before any chips go in. 

Palihapitiya wrote that it is “still the most obvious and fastest path to cash on cash returns.”  He added, “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.”

Palihapitiya said he and his partner, Anita Verma-Lallian, have locked in close to 6GW of power running through 2029. 

He had previously stated that zoning-approved land and silicon access hand their owners negotiating leverage over everyone downstream.

Why does Palihapitiya believe harnesses and applications will be the winners?

Above the concrete and the power lines, Palihapitiya’s pick is the harness. He stated in a post made in July, “A modern harness + open model will crush your token consumption but keep your performance.” 

A harness is the software wrapped around an AI model that decides what the model sees, which tools it can call, and when it stops, according to a Hugging Face glossary published on May 25. 

Anthropic’s Claude Code, OpenAI’s Codex, and Google’s Antigravity are all harnesses. Claude Code is referred to as “the agentic harness around Claude” in its documentation.

Palihapitiya stated that “the harness helps enterprises owns their proprietary context (what Alex Karp calls their ‘alpha’),” which to him includes their data, workflows, and business rules, among others. 

His thesis also stretches to applications, as he says they 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.”

Are people agreeing with Palihapitiya?

Some industry figures have chipped in their takes on Palihapitiya’s post, with many supporting it, especially his point on harnesses. 

Xiaoyin Qu, the founder of Tycoon AI, expressed more 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, in response to a different post that highlighted the performances of various AI agents, stated that the harness is “going to become the most important variable” in the AI stack, sitting right next to raw model capability.

In mid-July, Palihapitiya made a post on X that questioned the current state of AI spending, asking if it was paying off for anyone beyond the handful of firms already collecting the money, pointing to buyers who can now spend $0.50 per million leading-edge tokens instead of $56 for the same volume. 

It may seem that the harness call is his answer to his own complaint because if models get cheap and interchangeable, the money moves to whoever controls the data, the workflows, and eventually the applications sitting on top.

If you're reading this, you’re already ahead. Stay there with our newsletter.