In the week leading up to earnings, a certain level of anxiety develops around a stock like Palantir. Not the typical kind, where investors worry about an EPS miss of just one penny. The kind that occurs when a business is down 19% from the beginning of the year despite trading at 82 times sales and increasing revenue by 70% annually. On May 4, Palantir releases its Q1 results after the bell. The May 4 earnings test will reveal whether the air is gradually leaving this situation or simply resetting for a higher move.
By any reasonable software industry standard, the numbers that go into the print are extraordinary. Revenue for the fourth quarter was $1.41 billion, up 70% from the previous year. Commercial sales in the US increased by 137%. Government revenue in the US increased by 66%. In just the most recent quarter, the company closed contracts worth over $4.2 billion. This time, Wall Street anticipates $1.54 billion in revenue and about 28 cents in EPS, which would be a 74% increase in revenue. Guidance for the entire year 2026 ranges from $7.182 billion to $7.198 billion. During the most recent earnings call, CEO Alex Karp referred to Palantir as “an n of 1″—a term you can only use when the growth rate supports the swagger.
However, the truly intriguing aspect is that the stock has been moving against the fundamentals. Palantir reached $207 in November of last year. Now it’s $143. The shares are “overvalued today” at a price-to-sales ratio that surpassed 80x at the peak, according to Brett Schafer of The Motley Fool. Institutional investors believe that the AI software market overtook itself in the latter part of last year, with Palantir suffering the most after having run the hardest. The earnings narrative remained unchanged. The readiness to pay for it did.
The political climate is another factor, and it’s becoming noticeably colder. Campaigners in Minneapolis are putting pressure on the Swiss National Bank to sell its $1.1 billion stake in Palantir due to the company’s agreements with U.S. Immigration and Customs Enforcement. Never quiet, Peter Thiel’s profile has become more prominent. The argument used to be resolved by Alex Karp’s defense of the company’s federal work, which includes protections against government overreach. It’s no longer really resolving the issue. With all the ensuing concerns about surveillance, critics continue to point to the IRS partnership news that surfaced this week as another indication that Palantir is turning into the central component of the federal data infrastructure.
According to reports, Michael Burry, who shorted the housing market, has bet against Palantir, claiming that Anthropic and other private AI firms pose a threat to the enterprise software narrative. It remains to be seen if he is correct. What Palantir sells, Anthropic does not. Foundry and AIP are operating systems for institutional data, not language models. However, the more general point is made. Palantir’s moat, which is based on years of grinding through complicated government deployments, is more difficult to explain to a CFO comparing demos, and many well-funded companies are suddenly vying for the same enterprise AI dollars.
You get the impression that May 4 is truly binary as you watch this play out. John McPeake of Rosenblatt is betting on the same beat-and-raise pattern that has characterized Palantir’s recent quarters, with a $200 price target and a Buy rating. At $194.06, the TipRanks consensus suggests an upside of roughly 35%. But in that sentence, “consensus” is doing a lot of work. Another quarter of triple-digit U.S. commercial growth and non-softening language regarding the second half of the year are necessary for the bull case. Almost nothing is needed for the bear case—just a slight slowdown, a cautious tone from Karp, or an incorrect competitor name-check.
It’s difficult to ignore the fact that Palantir, a stock with a $342 billion market capitalization, trades more like a high-conviction thesis than an established software company. Investors appear to think that Palantir has a structural advantage in government and regulated industries that the hyperscalers can’t easily match, and that the AI infrastructure layer is a winner-take-most market. They may be correct. It’s possible that they are overpaying. The market will weigh in once more on May 4.
