Reading time: 7 min
Alex Karp stood in front of Wall Street’s best quarterly numbers in Palantir’s six-year history as a public company and watched his stock drop. Revenue up 85% year-on-year. US business doubling. Full-year guidance lifted by ten percentage points. Rule of 40 score at 145. And when the opening bell rang on May 5, 2026, PLTR opened at $140.28 — down $5.75 from the previous close — and kept falling, finishing the session off 7.08% at $135.68. That’s the Palantir paradox, still unresolved, and arguably more acute now than at any point since the company’s direct listing in 2020.
What the Numbers Actually Said
Palantir reported Q1 2026 revenue of $1.633 billion, beating the LSEG consensus of $1.54 billion by 5.9 percent. Adjusted earnings per share came in at $0.33, clearing the $0.28 estimate by 17.9 percent — the kind of beat that, for most software companies, would have triggered a meaningful after-hours rally. Net income nearly quadrupled to $870.5 million, or $0.34 per diluted share, from $214 million in the same quarter a year earlier. Adjusted operating income reached $984 million, a 60 percent margin. The company closed 206 deals valued at $1 million or more in the quarter, including 72 at more than $5 million, across a US commercial customer base that grew 42 percent year-on-year to 615 clients.
The headline the company wanted investors to focus on was US revenue. Total US revenue reached $1.28 billion in the quarter, up 104 percent year-on-year and 19 percent sequentially — the first time Palantir has crossed the 100 percent growth threshold in its home market since its direct public offering in September 2020. US commercial revenue rose 133 percent to $595 million. US government revenue grew 84 percent to $687 million, accelerating from the 66 percent recorded in the fourth quarter of 2025. Karp, writing in his shareholder letter, compared the company’s Rule of 40 score of 145 to a “feat matched only by NVIDIA, Micron, and SK Hynix.” Management lifted full-year 2026 revenue guidance to $7.65 billion to $7.662 billion, a 71 percent year-on-year increase and ten percentage points ahead of the February guidance, surpassing the $7.27 billion LSEG consensus. Second-quarter revenue guidance was set at $1.8 billion, above the $1.68 billion the Street had modelled.
Why It Didn’t Matter
Put that in context. The Nasdaq Composite closed May 5 up 1.03 percent to a fresh all-time high of 25,326.13. The broader software sector gained 2.31 percent on the session. Palantir fell 7.08 percent — a 9.4 percentage point underperformance against the sector it supposedly leads. The stock is down roughly 19.5 percent year-to-date through May 5, and still trading 34.8 percent below its 52-week high of $207.52, set in November 2025.
The problem is not the business. The problem is the price that was already attached to the business before Karp opened his mouth. Entering Monday’s close, PLTR’s trailing price-to-earnings ratio sat at approximately 227 times earnings, per MacroTrends data. The forward P/E, based on consensus 2026 earnings estimates, sits at 111 times, according to GuruFocus — compared to a software sector median of 18 times. Nvidia, which Karp invoked for comparison purposes, is growing at a similar 77 percent pace and trades at roughly 41 times earnings and 23 times sales, per analysis published by Gotrade. GuruFocus assigns PLTR a fair value of $63.34, which puts the current price 130 percent above its estimated intrinsic value. The company’s own cash, cash equivalents, and short-term Treasury securities totalled $8.0 billion at the end of March, per the Q1 filing — and it has paid no dividends and conducted no share repurchases.
This is the structural issue the earnings release could not resolve. Investors who own PLTR at current levels are not paying for what Palantir has done. They’re paying for a very specific version of what Palantir does next, and the market spent Tuesday recalibrating whether that version is still plausible at 111 times forward earnings.
The Competitive Argument That Didn’t Go Away
Valuation compression is being accelerated by a second concern: competitive moat. HSBC analyst Stephen Bersey downgraded PLTR to neutral from buy on May 4, lowering his price objective to $151, citing the emergence of agentic AI frameworks from rivals including Anthropic and OpenAI that he said challenge Palantir’s traditional differentiation. Palantir’s model relies heavily on embedding engineers directly into client operations to build bespoke data infrastructure — a model that generates high switching costs but is increasingly replicable, in Bersey’s view, as large language models become commoditised. “Traditional barriers to entry have begun to erode,” Bersey wrote in his note, warning that “the increasing proliferation of agentic AI frameworks increases the risk of multiple compression as the differentiator of Palantir narrows,” per reporting by FX Leaders. DA Davidson cut its price target from $180 to $165 the same morning. RBC Capital Markets analyst Rishi Jaluria, who carries the sole Sell rating on the stock with a $90 objective, had separately flagged possible customer attrition and the absence of capital returns despite a cash and Treasury securities balance of $8.0 billion as of March 31, per Palantir’s Q1 filing.
Cathie Wood’s ARK Invest has been adding to its PLTR position in recent months, and Wedbush’s Dan Ives maintains the Street’s highest target at $230 with an Outperform rating, arguing that Palantir is on a path to becoming a trillion-dollar AI company. The consensus across Wall Street sits at an average price target of roughly $194, implying 43 percent upside from the May 5 close — but that consensus has coexisted with a 20 percent decline in the share price over the same period the estimates were compiled. Consensus targets and stock performance have been pointing in opposite directions for most of 2026.
The Government Angle That Bulls Keep Coming Back To
The argument that most consistently holds up under scrutiny is Palantir’s structural position in US defense and intelligence spending. Trump’s 2026 defense budget request, at $1.5 trillion, is the largest in American history, and Palantir has built 23 years of embedded relationships across the Department of Defense, the intelligence community, and civilian agencies — relationships that are not easily displaced by a foundation model startup. The company disclosed a $300 million USDA contract win in the weeks before the earnings release, and US government revenue growing 84 percent year-on-year in Q1 — accelerating from 66 percent in Q4 2025 — confirms the thesis has not stalled. Unlike commercial software companies, which must re-sell their products to every new enterprise customer, Palantir’s government relationships function more like multi-decade infrastructure installations, with upgrade cycles driven by operational necessity rather than budget discretion. Oppenheimer initiated coverage with an Outperform rating and a $200 price target in late April, with analyst Param Singh arguing that the “adhesive nature of Palantir’s platform framework” creates durable revenue visibility that justifies a premium. The challenge for that premium, in the current environment, is that it needs to justify roughly 111 times forward earnings — and the broader AI investment landscape, where capital is flooding into competing platforms and enterprise AI spending is accelerating across every sector, offers no shortage of alternatives at cheaper multiples.
What the Market Is Telling You
Palantir has now delivered eleven consecutive quarters of accelerating revenue growth. It has reached GAAP profitability, generated nearly $1 billion in adjusted operating income in a single quarter, doubled its US revenue, and raised full-year guidance by ten percentage points above its own February forecast. And the stock is down 19.5 percent year-to-date. The market is not saying the business is broken. The market is saying that all of the above was already in the price at $160, at $170, at the November 2025 high of $207. The question now is whether 71 percent revenue growth — which would be a generational result for any software company operating at this scale — is enough to close the gap between where the stock is and where the multiples still imply it needs to go.
That is the same question AI-adjacent companies across the technology sector are beginning to confront as earnings season reaches its peak weeks: not whether AI is generating real revenue, but whether the revenue is compounding fast enough to justify the premiums the market assigned in 2024 and 2025 when the narrative was still hypothetical. For Palantir, the answer on May 5, 2026 was a 7 percent decline on a day when the Nasdaq hit an all-time high. The market is telling you something about what it thinks happens next — and it is not what Alex Karp’s shareholder letter said.
