Why Michael Burry is Betting Against Palantir's $300 Billion Valuation

Michael Burry, the investor famous for betting against the housing market before 2008, is making headlines again—this time targeting data analytics company Palantir. In late 2024, he took short positions against Palantir stock, arguing that its current market valuation is fundamentally disconnected from its business reality. To support his thesis, Michael Burry published a comprehensive essay exceeding 10,000 words on Substack, meticulously breaking down why he believes the stock is significantly overpriced.

His timing is striking. While the market has pushed Palantir’s valuation to nearly $300 billion—with Wall Street analysts on average rating it as overweight—Michael Burry sees a different picture. He believes the company’s AI ambitions mask deeper operational and financial concerns that the market has overlooked.

The $3.96 Billion Burden: Palantir’s History of Spending

Before Palantir went public in late 2020, the company operated as a well-connected Washington insider with strong relationships across government and Silicon Valley. Behind the scenes, however, the financial reality was starkly different. When Palantir filed its official S-1 registration statement in summer 2020, public disclosure revealed a sobering track record: the company had accumulated $3.96 billion in total losses as of June 30, 2020.

The losses weren’t ancient history. In just 2018 and 2019 combined, Palantir burned through $1.2 billion. During Series K funding in 2019, investors valued the company at $11.38 per share and committed $899 million. Yet between major financing rounds, the company relied on revolving credit lines to maintain cash flow—a pattern suggesting operational cash generation remained weak.

Perhaps most telling was the board’s decision in August 2020. Just before the direct listing, executives awarded CEO Alex Karp $1.1 billion in stock options. Michael Burry highlighted this moment as emblematic of the company’s relationship with capital: it knew how to deploy resources aggressively, whether toward operations or executive compensation.

Can AI Really Deliver? The Reliability Question

Palantir’s narrative transformed in 2023 when it launched its Artificial Intelligence Platform, positioning itself as a bridge between third-party large language models from OpenAI and Anthropic and enterprise customer data. Since then, the market has rewarded the AI pivot handsomely. Last year, Palantir reported $4.5 billion in annual revenue—a 56% increase from the prior year. The stock itself soared roughly 450% over the past two years.

But Michael Burry challenges the underlying AI thesis. He argues that Palantir’s dependence on third-party language models introduces a critical vulnerability: these models are “systematically unreliable” for mission-critical applications. To support this, he cited a Stanford University research paper documenting reasoning failures in large language models—failures that matter enormously for legal reasoning, scientific analysis, medical decision support, military targeting, and other applications demanding 100% precision.

This isn’t theoretical concern. If AI models lack the precision required for high-stakes decisions, their value in Palantir’s enterprise software diminishes substantially. Michael Burry sees an AI platform built on unstable technological foundations.

The Growth Disparity That Reveals the Real Business Model

When Michael Burry examined Palantir’s regional performance, a striking asymmetry emerged. U.S. commercial revenue jumped 137% in the most recent year. International commercial revenue, by contrast, grew only 2%. This dramatic divergence suggests something important about how Palantir’s business actually operates.

Michael Burry argues the numbers reveal Palantir functions more like a high-touch consulting firm than a scalable software-as-a-service platform. The steep U.S. growth likely depends on intensive engineer involvement and relationship-building on the ground. International stagnation signals that this model doesn’t travel well or replicate easily.

If that interpretation is correct, Palantir faces significant constraints. Well-funded competitors like Salesforce and Microsoft could eventually decide to move aggressively into data integration. Once savvy customers recognize that “Emperor Palantir has no clothes”—or once AI tools make data integration cheap and accessible—the competitive moat collapses. Michael Burry believes that inflection point is approaching.

The $100 Billion Question

Michael Burry’s final argument is perhaps the boldest. He forecasts that recent momentum will not persist. The company’s winning streak in the AI era will prove temporary. When the market eventually reprices Palantir to reflect the underlying business dynamics, he predicts, the true valuation will settle below $100 billion.

That represents a dramatic repricing from the current $300 billion market cap. For Michael Burry, the risk-reward calculus is compelling. He is positioned to profit significantly if his thesis plays out. Meanwhile, the broader market continues to price Palantir as a stable, high-growth AI beneficiary—a fundamental disagreement that will eventually be resolved.