Palantir (PLTR.US) U.S. commercial revenue surges 148%, yet Wall Street is debating whether the "AI application myth" can continue

Palantir Technologies (PLTR.US), the AI application leader focused on "AI + data analytics," reported its results on Tuesday morning Beijing time, and two key facts held true simultaneously: U.S. commercial revenue surpassed U.S. government revenue for the first time, proving that AI software is monetizing at scale on the enterprise side; however, on the other hand, the year-over-year growth rate of total value of new U.S. commercial contracts signed fell from over 153% last quarter to 11%, creating a sharp divide among institutional and retail investors over how far they can see Palantir's next phase of AI application growth amid the AI agent frenzy. That is why, despite both results and guidance beating Wall Street analysts' consensus expectations, the stock still fell as much as about 10% in U.S. pre-market trading. Regarding the latest results from AI application leader Palantir and management's optimistic outlook, the core judgment of Morgan Stanley's latest research note is that Palantir is a major beneficiary of enterprise "sovereign AI" and the spread of open-weight models, highlighting that third-quarter revenue growth accelerated for the 13th consecutive quarter and that the adjusted operating margin reached 61.7%; at the same time, multiple order metrics for customer additions slowed. Morgan Stanley therefore took an action that appears contradictory but is in fact consistent: it sharply raised 2027 and 2028 revenue and earnings forecasts and lifted its price target from $205 to $233, while maintaining "Equal-weight"a relatively cautious, neutral-leaning ratingrather than upgrading to "Overweight," the most bullish rating. Based on the closing price of $178.34, $233 implies 30.65% potential upside. After surging 340% in 2025, Palantir's stock has cumulatively risen more than 130% over the course of 2025, making it a "bull market myth of the AI application software sector," and some veteran Wall Street analysts, including those at Morgan Stanley, believe the stock still has room to rise. In October just past, Palantir's stock fell from $187.05 to $184.11, a monthly decline of about 1.57%, outperforming the Nasdaq 100 Index and the Philadelphia Semiconductor Index by about 5.08 and 13.79 percentage points, respectively. But from the intraday stage high of $212.94 on October 12 to the pre-earnings close of $178.34 on November 2, the stock had already pulled back about 16%, failing to continue an independent rally as the AI trade retreated. U.S. commercial revenue surpassed government business for the first timewhy can nearly doubled revenue not dispel order doubts? Palantir's earnings data show that third-quarter revenue rose from $1.181 billion a year earlier to $2.313 billion, up 96% year over year and 20% quarter over quarter, above the consensus estimate of $2.18 billion; adjusted earnings per share were $0.44, above the consensus estimate of $0.41. The strongest growth remained in the United States: commercial revenue was $984 million, up 148% year over year and 29% quarter over quarter, surpassing U.S. government revenue of about $953 million for the first time; the latter grew 96% year over year, also above last quarter's 90%. By comparison, international commercial and government revenue grew 21% and 31%, respectively, without the simultaneous explosion seen in the U.S. market. Profit was also cash-backed: GAAP operating income was $1.085 billion, a 47% margin; adjusted operating income was $1.427 billion, a margin of about 62%. On the company's basis, the "Rule of 40" scorerevenue growth plus adjusted operating marginreached 158%. Operating cash flow was $1.205 billion, and free cash flow after capital expenditures was about $1.197 billion; the company's adjusted free cash flow was $1.216 billion. The "Rule of 40" is a rule-of-thumb metric for whether a software company can balance growth and profitability. On Palantir's basis, adding revenue year-over-year growth to adjusted operating margin, a total of 40% counts as meeting the standard. Palantir's third-quarter revenue grew about 96% year over year, meaning revenue was nearly double the year-earlier level; adjusted operating margin was about 62%, meaning every $100 of revenue left about $62 of adjusted operating profit. The two add up to about 158%, far above the 40% reference line. The 158% here is a composite score, not a margin, nor a profit growth rate. Guidance and management's outlook are similarly strong. Fourth-quarter revenue is expected at $2.6492.653 billion, with the midpoint about 7.8% above the LSEG-compiled analyst consensus of $2.46 billion; full-year revenue guidance was raised to $8.5308.534 billion, implying growth rising from about 82% to 91%, U.S. commercial revenue is expected to exceed $3.559 billion, up at least 143%, and full-year adjusted free cash flow is expected at $4.85.0 billion. Morgan Stanley calculates that if fourth-quarter revenue comes in another 5% above the guidance midpoint, year-over-year growth could approach 98%, but this is Morgan Stanley's scenario estimate, not a company commitment. What sell-side and buy-side are really debating is orders after revenue. Total contract value (TCV) of new contracts signed company-wide this quarter was $2.918 billion, up only 6% year over year; U.S. commercial TCV was $1.463 billion, up only 11% year over year. Remaining performance obligations (RPO) at period end were about $4.9 billion, up 88% year over year but flat quarter over quarter. Morgan Stanley also noted that net U.S. commercial customer additions fell from 38 to 33. Management explained on the call that several sovereign AI initial contracts worth tens of millions of dollars had shorter-than-usual terms; shorter terms themselves the full-term contract amount, and contracts with original terms of no more than 12 months are not included in the company's disclosed RPO. At the same time, net dollar retention rose from 157% to 165%, and U.S. commercial TCV signed over the past 12 months still grew 62%. Therefore, the Morgan Stanley analyst team stressed that the sharp drop in single-quarter order growth cannot be directly interpreted as a collapse in demand; but the market needs to see this batch of short-term contracts renewed and expanded on schedule to confirm how much of today's revenue growth has a chance to continue. From model capability to customers' AI workflow budgets: why does Morgan Stanley raise cash flow but cut the valuation multiple? The key to Palantir benefiting from the agent wave is that it can connect models into the processes enterprises actually run. Traditional data warehouses can calculate which orders may be delayed; Palantir's Ontology further links orders, inventory, factories, and personnel permissions, enabling agents to propose allocation plans within authorized limits, trigger approvals, and test decisions against execution results. The "sovereign AI" Morgan Stanley highlights in the earnings is also not limited to government procurement: enterprises can retain control over proprietary data and model weights, post-train and evaluate open-weight models through AIP Evolve, and choose models with the right cost and performance for each task. Two specific examples from the management call cited by Morgan Stanley show that L3Harris completed fine-tuning for a specific task in less than 48 hours, with initial test performance exceeding the frontier model previously used and costs reduced by 95%; a global home appliance company started with a $1 million pilot and by the third quarter had expanded to a $15 million annual contract value, covering commercial, finance, and supply chain processes. The former shows why customers adopt, and the latter shows how Palantir earns more revenue from adoption. This strong expansion capability is the main basis for Morgan Stanley's model upgrades. The bank raised its 2027 and 2028 revenue forecasts by 19% and 28%, respectively, to $14.08 billion and $21.14 billion, implying year-over-year growth of 65% and 50%; among them, U.S. commercial revenue forecasts were raised to $6.80 billion and $11.42 billion. Over the same period, adjusted operating margin was raised from prior estimates of 61.0% and 62.1% to 63.5% and 65.5%, adjusted earnings per share from $2.31 and $3.27 to $2.79 and $4.31, and free cash flow in the model rose to about $8.39 billion and $12.84 billion. This also explains the stock's resilience relative to chip stocks: Palantir can sell more and larger-scale workflows to the same customers and already generates substantial cash flow, without first building a new data center to have a chance at the next software revenue dollar. Morgan Stanley raised its price target but maintained "Equal-weight" because it is bullish on the growth opportunity from sovereign AI but remains reserved about whether a large customer base can sustain the current pace of expansion over the long term. Morgan Stanley's research found that some large customers spending more than $5 million remain sensitive to price and total deployment cost. Average revenue from the top 20 customers over the past 12 months has reached $146 million, while overall customer count growth slowed from 24% to 20%; the more growth depends on existing customers buying more, the more it depends on whether new use cases can repeatedly prove returns. Morgan Stanley's $233 base-case target already assumes 37% revenue CAGR over the next five years, reaching $41.2 billion in 2031, with free cash flow of $27.5 billion and a margin of about 67%. On valuation, Morgan Stanley changed its prior "2030 free cash flow per share of $7.07, at 41x" to "2031 free cash flow per share of $10.98, at 30x," then discounted at a 13% weighted average cost of capital and added back $7.9 per share of net cash. The higher price target comes from higher cash flow forecasts, not from giving the stock a richer multiplewhich also means the reason Morgan Stanley's price target can still rise is that Morgan Stanley expects Palantir to earn more cash in the future, while retaining a measure of caution about how quickly it can continue to expand purchases from existing customers. If new customers and per-customer spending expand faster, Morgan Stanley's bull case is $336; if deployment costs limit adoption, the bear case is just $91, representing 88.4% upside and 49.0% downside from the base-case closing price, respectively. Wall Street is deeply divided: why can the same earnings report support both $265 and $130? After the earnings, Citigroup maintained its "Buy" rating on Palantir and raised its price target from $250 to $265, implying about 48.6% upside over the next 12 months. Citigroup's analyst team focused more on the simultaneous strength of U.S. commercial and government business and believes contract terms affected the comparability of order figures. UBS Group AG maintained "Buy" and Baird maintained "Outperform," both with $250 price targets, implying 40.2% upside; UBS Group AG noted that U.S. commercial revenue growth exceeded its estimate range, but flat RPO quarter over quarter was weaker than its quantitative model expected. Piper Sandler maintained "Overweight" and raised its price target from $231 to $235, implying 31.8% potential upside. On the cautious side, Cantor Fitzgerald maintained "Neutral" and a $207 price target, implying 16.1% potential upside; Royal Bank of Canada Capital Markets sharply raised its price target from $90 to $130 but maintained "Underperform," implying 27.1% pessimistic downside. These ratings and price targets show that the real disagreement between Wall Street bulls and bears is not whether AI software has customers, but whether the pace of expansion among existing customers can last long enough to support today's valuation. Therefore, the most telling signal next quarter will be whether U.S. commercial new contract value can rebound while net dollar retention and cash margins remain strong.