DENVER, November 3, 2026 — Palantir Technologies reported a 96% year-on-year increase in third-quarter revenue to $2.31 billion and raised its full-year outlook, as demand for its artificial-intelligence software pushed US commercial sales above revenue from US government customers for the first time. The results were stronger than analysts expected, but investors also focused on a sharp slowdown in the growth of new commercial contract value and the company’s high valuation.

The software company said revenue reached $2.313 billion in the three months ended September 30, compared with $1.181 billion a year earlier. US commercial revenue rose 148% to $984 million, while US government revenue increased 96% to $953 million. Net income attributable to common stockholders was about $1.13 billion, and diluted earnings per share were 44 cents, according to Palantir’s quarterly business update.

Palantir now expects 2026 revenue of approximately $8.53 billion to $8.534 billion and fourth-quarter revenue of $2.649 billion to $2.653 billion. Both ranges were above analyst estimates cited by Reuters. The guidance suggests the company expects demand to continue into the final quarter, even as the market weighs whether its extraordinary growth can be sustained and whether the share price already reflects much of that optimism.

The most notable shift in the quarter was the crossover between commercial and government revenue in the United States. Palantir has long been associated with defence, intelligence and public-sector data systems, but the latest figures show that private-sector customers are now a larger source of US revenue. Commercial sales rose to $984 million, up from $397 million in the same quarter a year earlier, while government revenue reached $953 million.

The company’s business update said total US revenue increased 119% year on year to $1.94 billion. International revenue grew more slowly, at 26%, according to reporting by Barron’s. That gap matters because it shows how much of the current acceleration is concentrated in the US market, where companies and public agencies are moving quickly to evaluate AI tools. Strong domestic growth does not guarantee that the same pace will be repeated in regions with different procurement rules, budgets and attitudes toward data sovereignty.

Palantir sells software that helps organisations combine data from multiple systems, model operational processes and use AI tools within those workflows. Its Artificial Intelligence Platform, or AIP, is designed to connect large language models with an organisation’s own data and processes. The company argues that its software lets customers retain control of their data and models while applying AI to decisions in manufacturing, logistics, healthcare, finance and government operations.

Those capabilities address a practical barrier to enterprise AI adoption. Many organisations can test a model in a small demonstration, but putting it into production requires permissions, reliable data, integration with existing systems and a way to monitor what the model does. Palantir’s software is built around those operational requirements. Whether the value it creates justifies its price depends on the results each customer achieves and how much of the work could eventually be performed by other software platforms.

Palantir reported that it closed 212 deals worth at least $1 million during the quarter, including 107 deals of $5 million or more and 64 of $10 million or more. It also reported a net dollar retention rate of 165%, an increase of eight percentage points from the previous quarter. Retention measures whether existing customers are spending more or less over time; a high figure can indicate expansion within accounts, although it does not by itself show how profitable every contract is.

The company’s investor presentation highlighted examples in which customers used its software to support supply-chain planning, chemical manufacturing, aviation safety and AI deployments that keep data within customer-controlled environments. Those examples are Palantir’s own descriptions of customer projects and should be treated as company-reported outcomes, not independent audits of every claimed saving or performance improvement.

Palantir’s revenue growth and guidance beat expectations, yet its shares fell in after-hours trading after the results. Reuters and Barron’s reported declines of around 3% to nearly 4%, with the exact move varying during the trading session. The reaction reflected the high expectations built into the stock as much as the operating results themselves. A company can deliver a strong quarter and still disappoint investors if they believe the next phase of growth will be slower or more expensive.

The main concern was the growth of new commercial contract value. Reuters reported that US commercial total contract value growth slowed to about 11% year on year, from more than 153% in the previous quarter. Palantir said it signed several large but shorter-than-usual deals associated with early deployments of sovereign AI. The company’s reported revenue can continue rising as existing contracts are recognised, even when the value of newly signed business grows more slowly.

Contract value, revenue and remaining performance obligations measure different things. Revenue is recognised as a company delivers services. Contract value can include the amount associated with new agreements and options, while remaining performance obligations represent contracted revenue that has not yet been recognised, subject to the company’s accounting definitions. A slowdown in one measure does not automatically predict a decline in another, but investors look at them together to judge how much future growth is already supported by signed business.

Palantir’s guidance implies that the company expects to maintain a high level of activity in the fourth quarter. But guidance is a forecast, not a guarantee. The final result will depend on contract timing, customer deployment schedules, the pace at which new projects move from pilots to production and the company’s ability to expand outside its strongest markets. Large contracts can also be lumpy: one deal signed earlier or later than expected can move a quarter’s total materially.

Valuation adds another layer. Before the earnings report, Palantir traded at a premium multiple of projected earnings, according to Barron’s, reflecting investor confidence in the company’s AI opportunity. At such a valuation, the market may demand not only strong revenue growth but evidence that growth can persist over several years. If contract momentum slows, international growth remains modest or competitors offer similar capabilities at lower cost, the stock can come under pressure even if the company remains profitable.

Also read: Alphabet Shares Fall 12.6% as AI Spending Weighs on Investors

Palantir’s public-sector business remains a significant part of its identity. Government customers use data systems for a range of activities, including defence planning, logistics, intelligence analysis and operational coordination. These contracts can be large and long-lived, but they can also be exposed to budget changes, procurement disputes and political scrutiny. The company’s growing commercial business may diversify its revenue base, yet government work continues to shape how customers and investors understand its products.

The company has also promoted “sovereign AI”, a model in which organisations keep greater control over the data, infrastructure and models used in their AI systems. That pitch can appeal to governments, defence organisations and companies operating under strict data-protection rules. A customer may want the benefits of advanced models without sending sensitive information to a general-purpose service or relying entirely on a vendor’s cloud environment.

Sovereign deployments can involve trade-offs. Keeping data and models under an organisation’s control may improve governance, but it can require more technical staff, infrastructure and ongoing maintenance. A system still needs to be tested for accuracy, security and reliability, and its outputs must fit the organisation’s actual workflow. The label “sovereign” does not by itself guarantee privacy or performance; those depend on architecture, permissions, contracts and operational controls.

Competition is another issue. Cloud providers and model developers are building enterprise products that connect AI with business data, while specialist software companies are adding AI features to existing tools. Customers may choose a platform that is easier to deploy, already integrated with their systems or priced more predictably. Palantir’s advantage will depend on whether its software delivers measurable improvements that are difficult to reproduce with competing products.

There are also public-interest questions when software is used by police, immigration agencies, militaries or other institutions with significant authority over people. Data integration can improve coordination, but it can also make it easier to combine records in ways that affect privacy, due process and accountability. The relevant questions are what data is collected, who can access it, how decisions are reviewed and whether affected people can challenge errors. Commercial growth does not remove the need for those safeguards.

Also read: Anthropic’s $518 Billion AI Infrastructure Bet Locks In Huge Compute Commitments

The next results will show whether Palantir can convert its current pipeline into sustained revenue growth without relying on a small number of very large contracts. Investors will watch the growth of US commercial revenue, the value and duration of new contracts, international expansion and the company’s ability to keep margins high while it hires staff and invests in sales and product development.

Palantir’s business update reported adjusted operating income of about $1.43 billion, a 62% adjusted operating margin, and adjusted free cash flow of approximately $1.22 billion. Those measures indicate the company is generating substantial cash from its current scale. They are non-GAAP measures, however, and the company’s own presentation cautions investors to consider them alongside the corresponding figures prepared under generally accepted accounting principles. Stock-based compensation and other adjustments can materially affect comparisons.

Customer retention will be another signal. A 165% net dollar retention rate suggests existing customers are expanding their use of Palantir’s software, but the rate can change if large customers finish a deployment, reduce usage or shift to another provider. The number of customers and the distribution of revenue across them also matter. A broad base of growing accounts can be more durable than expansion driven by a few unusually large customers.

International growth will test whether Palantir’s model travels easily beyond the United States. Government and enterprise buyers in Europe and Asia may have different rules for data residency, public procurement and AI governance. Some may prefer local vendors or require systems to run on infrastructure they control. Palantir’s sovereign AI pitch could help address those concerns, but selling and supporting deployments across multiple regulatory environments can increase costs and lengthen deal cycles.

The company’s latest results show that AI spending is translating into significant revenue for at least one software provider. They do not establish that every AI application will produce similar returns or that the current pace of growth will continue indefinitely. The important test is whether customers keep expanding their use of the platform, whether new contracts support future revenue and whether the company can defend its margins as competitors move into the same market.

For now, Palantir has delivered a record quarter and raised its outlook, with US commercial business becoming its largest US revenue segment. The market’s caution reflects the gap between strong reported results and the even higher expectations embedded in the stock. The next few quarters will help determine whether the company can sustain its pace, broaden growth beyond the United States and turn early AI adoption into recurring, profitable deployments.