Michael Burry built his reputation by being early and right about a bubble nobody else could see.

On Wednesday he turned that instinct on Palantir Technologies (PLTR), calling the company an AI consultant riding a wave of corporate FOMO and warning its market value could eventually fall below $100 billion.

By the time his post went up, Palantir had already banked a new Army contract two days earlier, and within hours PwC piled on with an expanded alliance of its own.

That timing is the story. Bear cases usually land in a vacuum, giving the market time to sit with the discomfort. This one landed in the middle of two corporate wins, and the stock barely flinched.

Burry’s bear case rests on the balance sheet

Burry’s argument, laid out in a Seeking Alpha writeup of his post, is not really about growth. It is about what kind of company Palantir actually is.

He compared Palantir’s deferred revenue ratio, about 32%, to Accenture’s roughly 31% and to subscription software firms like Salesforce and ServiceNow, which run between 80% and 207%.

Related: Jim Cramer explains Palantir, Salesforce rebound

The implication is that Palantir bills and books revenue more like a consulting shop than a software platform, which matters because consultants trade at far lower multiples than SaaS companies do.

He also flagged accounts receivable climbing to $1.49 billion as of June 30 from $1.04 billion at the end of 2025, with one customer responsible for 27% of that balance despite no single customer accounting for more than 10% of revenue.

Rising receivables tied to a concentrated customer can signal that a company is booking revenue faster than it is actually collecting cash.

A few additional details rounded out the critique:

  • Palantir reported about $1.6 billion in pretax GAAP income in 2025 but paid no federal cash taxes, a gap Burry linked to stock-compensation deductions that pushed federal net operating loss carryforwards up to $9 billion.
  • The company canceled a $1 billion buyback authorization after repurchasing only about $75 million of stock in 2025.
  • CEO Alex Karp’s personally owned aircraft cost the company $17.2 million in 2025, more than double the $7.7 million spent a year earlier.
Palantir shares rose nearly 7% after a PwC alliance and Army contract offset Michael Burry’s renewed $100 billion valuation warning.

JHVEPhoto / Getty Images

PwC just expanded its Palantir alliance

Hours after Burry’s post, PwC announced it was deepening its alliance with Palantir to build what the firms called the industry’s first AI-native deals platform, running on Palantir’s Foundry and AIP software, according to a press release from PwC.

The platform targets mergers, acquisitions and divestitures, and the firms say it is designed to help clients execute deals up to 50% faster while cutting one-time transaction costs by as much as 45%.

That matters because it is enterprise validation, not government spending. Consulting giants do not attach their name to a platform unless they expect client demand to follow, and PwC was recently named a leader in Palantir’s own ecosystem for AI engineering work.

Shares climbed nearly 7% to around $181 on the news, according to Benzinga, even as Burry’s fresh short thesis was circulating the same morning.

The Army handed Palantir eight new systems

Two days before any of that, the Army Contracting Command awarded Palantir USG, a wholly owned subsidiary, a prime agreement to produce eight TITAN ground station systems, according to a press release distributed on the company’s behalf.

TITAN is a crewed AI-enabled ground station that pulls in space, aerial and terrestrial sensor data to support targeting and long-range fires, and the award covers four Advanced and four Basic variants built with partners including Anduril and L3Harris.

That contract matters because it extends Palantir beyond a single prototype into sustained production, the kind of multi-year commitment that defense investors weight more heavily than commercial pilots.

It is also the exact kind of government revenue Burry’s thesis does not directly address. His numbers focus on receivables and tax treatment, not on whether the Pentagon still wants the product.

More Palantir:

Two narratives about Palantir are now running in parallel

None of this makes Burry wrong. Deferred revenue ratios and receivables concentration are real accounting signals, and they do not disappear because a stock rallies on unrelated news.

What changed this week is that Palantir now has three separate stories competing for the market’s attention within 48 hours, a bear case built on the books and two wins built on demand.

That pattern is becoming familiar across AI-adjacent software, where valuation skeptics and revenue catalysts increasingly arrive in the same week rather than the same earnings cycle.

Investors watching Palantir are no longer just betting on the business. They are betting on which narrative moves faster, and this week the demand story got the head start.

Related: Palantir CEO just made bet that could reshape defense-tech race