"Anthropic Is Eating Palantir's Lunch" — But Is Burry Right?
Michael Burry doesn't do subtle. The investor who called the 2008 housing crash — and whose bets were immortalized in The Big Short — posted a blunt verdict on X this week: "Anthropic is eating Palantir's lunch."
The timing was deliberate. Anthropic had just announced its revenue run rate crossed $30 billion — up from $9 billion at the end of 2025. Palantir's stock dropped 6% the same day. And Burry, who has held put options on Palantir since mid-2025, added fresh ammunition to a bear case he has been building for over a year.
But is the comparison actually fair? The answer requires understanding what these two companies are — and more importantly, what they are not.
The numbers that triggered Burry
Anthropic confirmed its revenue run rate has topped $30 billion, up from $9 billion at the end of 2025, with more than 1,000 business customers now spending over $1 million annually — a figure that has more than doubled since February. Bloomberg
To put that in context: Anthropic was at roughly $1 billion ARR in December 2024. It crossed $9 billion by end of 2025. It is now at $30 billion. That is 30x growth in 15 months. SaaStr No enterprise software company in history has scaled this fast.
 |
| Source: Bloomberg, Anthropic (Apr 2026), Ramp AI Index (Mar 2026). Reconstructed via AI for clarity. |
Burry's post was direct: "Anthropic went from $9B to $30B in months. It took $PLTR 20 years to get to $5 billion. Anthropic is taking 73% of all new enterprise spending per Ramp." Substack
That 73% figure comes from Ramp's March 2026 AI Index — a dataset tracking actual business spending across thousands of companies. Nearly one in four businesses on Ramp now pays for Anthropic, compared to one in 25 a year ago. Anthropic's adoption grew 4.9 percentage points month-over-month, its fastest pace yet, and it now wins about 70% of first-time, head-to-head enterprise purchasing decisions versus OpenAI. Benzinga
Those are not vanity metrics. Those are purchasing decisions.
What Burry's short position actually looks like
Before treating Burry's post as a market call, it helps to understand his position mechanics. Burry previously disclosed a sizeable short in Palantir via long-dated $50 strike put options expiring in 2027, giving him downside exposure to roughly 5 million shares. Benzinga His actual out-of-pocket cost was approximately $9.2 million in premiums — not the $912 million notional figure that made headlines last year.
This is a long-duration macro bet, not a day trade. Burry is not saying Palantir collapses this quarter. He is saying that at current valuations — which have Palantir trading at triple-digit earnings multiples — the stock is pricing in a future that the fundamentals do not support. The Anthropic post is fresh evidence for a thesis he has held for months.
The core misunderstanding: these are not the same business
Here is where the analysis gets more nuanced than Burry's X post suggests.
Anthropic builds the engine. Palantir builds the vehicle.
Anthropic is a foundation model company. Its revenue comes from businesses accessing Claude through APIs, enterprise subscriptions, and developer tools. Claude Code alone — Anthropic's agentic coding tool — now has a run-rate revenue above $2.5 billion, a product that went from zero to that figure in roughly nine months. SaaStr The model is the product. The faster and cheaper it gets, the more it sells.
Palantir is something structurally different. Its core platforms — Gotham for government and defense, Foundry for commercial enterprise — are data integration and decision-support systems. Palantir does not compete with Anthropic by building its own LLMs. It competes by helping organizations deploy LLMs — including Claude — safely, compliantly, and at scale within complex operational environments.
Palantir's AIP (AI Platform) is explicitly designed as a bridge: it allows enterprises to take models from Anthropic, OpenAI, or others and apply them to real business workflows with proper data governance, security controls, and audit trails. In that sense, Anthropic's growth is not purely a threat to Palantir. It is also a tailwind — more powerful models mean more demand for platforms that can deploy them responsibly.
Where Burry's bear case has real teeth
That said, dismissing Burry entirely would be a mistake. His structural critique of Palantir's valuation is harder to argue with than his Anthropic comparison.
Palantir's government contracts — the foundation of its business for two decades — are, as Burry notes, relatively low-margin and operationally intensive. The company historically deployed engineers directly into client environments to make its software work, a model that does not scale like a pure SaaS business. The commercial pivot is real, but it is still early.
More importantly, Palantir's stock price has been pricing in perfection. At 200-plus times earnings, there is no margin for error. If the commercial segment disappoints, or if enterprise clients decide that lighter-weight AI solutions — Claude via API, for instance — are sufficient for their needs without Palantir's platform layer, the valuation re-rates sharply downward.
That is the legitimate version of Burry's argument. Not that Anthropic makes Palantir worthless — but that Anthropic's success reduces the addressable premium that Palantir can justify charging.
Two possible futures
The market will ultimately resolve this debate in one of two ways.
In the first scenario, enterprise AI adoption continues accelerating and organizations find that raw model access is sufficient. They build internal workflows directly on top of Claude or GPT-4, bypassing the need for a Palantir integration layer. In this world, Burry is right and Palantir's commercial growth story stalls.
In the second scenario, AI deployment complexity scales with AI capability. As models become more powerful, the governance, security, and integration challenges multiply. Regulated industries — finance, healthcare, defense — cannot simply plug Claude into their systems without a robust oversight platform. Palantir's value proposition strengthens as AI gets harder to deploy safely, not easier.
Both scenarios are plausible. The data right now favors Anthropic's momentum. The structural argument about enterprise complexity favors Palantir's long-term position.
The bottom line
Burry's headline is provocative but incomplete. Anthropic's growth is genuinely extraordinary — $30 billion ARR, ahead of OpenAI, projecting positive free cash flow by 2027 while OpenAI pushes its breakeven to 2030. The-ai-corner The enterprise AI spending shift is real and it is accelerating.
But Palantir and Anthropic are not direct substitutes. One sells the intelligence layer. The other sells the operational infrastructure to deploy it. The question is whether that infrastructure layer remains essential or becomes commoditized as AI tools grow simpler.
Burry is betting on commoditization. Palantir is betting on complexity. The next twelve months of enterprise AI deployment data will tell us who is right.
In the meantime, Palantir's valuation leaves very little room to be wrong.
