Last week we talked about “the inference inflection.”

Seagate had just told us its top three cloud customers had nearly doubled their remaining performance obligations to $1.1 trillion, and that storage was sold out through 2027.

It was a strong clue that the hyperscalers would guide capex higher when they reported.

They did.

As we discussed, they are spending whatever it takes to build the infrastructure.

Because more supply (more compute) equals more revenue. 

Now the demand side is proving out.

Yesterday Palantir reported the strongest quarter in its history. Revenue grew 85% year-over-year, the fastest pace as a public company. U.S. Commercial revenue grew 133%. U.S. Government grew 84%. And the full-year guidance was raised to 71% growth, ten percentage points above the prior guide.

And this is the graphic that tells the story about the state of AI.

Palantir is now in rare company with Nvidia, Micron and SK hynix (of South Korea) as a rule of 145+ company (a measure of revenue growth + operating margin).  

AI has made each of those four companies faster growing and more profitable — with no end in sight.

So, Palantir is now positioned alongside the chip and memory companies that build the AI revolution — which implies the AI software application layer is part of AI infrastructure.

This is the big takeaway: Agentic AI (which Palantir is building for enterprises) is consuming compute the way data centers consume power.

It’s always on, and always inferencing — the “inference inflection.”

We’ve been building the AI-Innovation Portfolio around this thesis since June 2023. Endless demand drives outsized value at the companies that provide the scarce infrastructure resources of this fourth industrial revolution. Join us here, or click below for more details on how we’re positioned and why.