Key Highlights

  • Q1 2026 (ended ) delivered $1.633 billion of revenue, up 85% year over year, with US revenue up 104% and US commercial revenue up 133%.
  • Palantir launched a deployment engine for Nvidia's Nemotron open models in sovereign, classified and air-gapped environments, a confirmed launch, but with no named customers or disclosed contract value.
  • Shares closed at $131.53, near $315 billion of market value and roughly 60 times trailing sales, down about 17% over 12 months despite accelerating growth.
  • Concentration comes first: US government work supplied 42% of Q1 revenue, exposing Palantir to appropriations cycles and political turnover.

Palantir (NYSE:PLTR) pairs Nvidia's Nemotron open models with its own Ontology for classified and air-gapped deployments, while first-quarter US revenue grew 104%. Investors have marked the shares down anyway. Second-quarter results land 3 August.

A Company Growing Faster Than Its Share Price

Palantir Technologies Inc. (NASDAQ: PLTR) has spent 2026 in an unusual position: operating results keep improving while the stock does not. It is compounding revenue at a rate almost unheard of at its scale, has become central to US defence software, and in late June announced a joint engine with Nvidia for classified AI deployments. Yet the shares closed at $131.53, down roughly 17% over 12 months, according to aggregated market data.

That gap is the story heading into second-quarter results on 3 August. Investors are weighing whether US commercial demand is still compounding, whether the sovereign-AI push converts into contracted revenue, and whether any growth rate justifies the market's multiple.

What Palantir and Nvidia Have Actually Agreed

The Nvidia relationship has two distinct stages.

The first came, when the two said they would build an integrated stack, folding Nvidia's CUDA-X libraries, its open Nemotron model family, its cuOpt optimisation software and its Blackwell architecture into Palantir's Ontology and AI Platform (AIP). Nvidia named Lowe's among the first customers, for supply-chain modelling, but hedged: many described features "remain in various stages and will be offered on a when-and-if-available basis."

The second stage, matters most to the current debate. Palantir announced an engine for deploying Nemotron open models in sovereign environments, pairing Nvidia's models, NVIDIA AI Enterprise and NIM microservices with Palantir's AIP, Ontology, Foundry and Apollo, aimed at US government agencies and critical-infrastructure operators. Palantir describes four workstreams: deployment engineering for classified and air-gapped networks, context engineering, model engineering that modifies weights using customer-proprietary data, and enterprise-grade delivery. Customers retain data-authorisation controls, rights to erasure and auditability. Chief executive Alex Karp framed it as letting agencies "unleash the full power of LLMs while removing underlying security risks"; Nvidia's Jensen Huang said open-source AI "is foundational to national security."

What is confirmed: the launch, the product integration and the joint engineering commitment. What is not: any named agency customer, contract value, deployment timeline or performance benchmark. Treat the revenue impact as unquantified.

The Quarter That Reset Expectations

Palantir's Q1 2026 results, reported for the period ended were the strongest in its history as a public company.

Revenue reached $1.633 billion, up 85% year over year and 16% sequentially. US revenue of $1.282 billion grew 104%; within that, US commercial revenue of $595 million rose 133% and US government revenue of $687 million rose 84%. Net dollar retention was 150%, US commercial customers reached 571 (up 42%) and total customers grew 31%.

Bookings matched. Total contract value closed was $2.41 billion, up 61%, including $1.176 billion of US commercial TCV. US commercial remaining deal value, contracted but unrecognised backlog, reached $4.92 billion, up 112%.

GAAP operating income was $754 million (46% margin) and GAAP net income $871 million. Adjusted operating income, a non-GAAP measure, was $984 million for a 60% margin, producing a "Rule of 40" score of 145%. GAAP diluted EPS were $0.34, adjusted diluted EPS $0.33, and adjusted free cash flow $925 million on a 57% margin.

Management then raised full-year guidance by what chief financial officer Dave Glazer called the largest margin in company history, about ten percentage points. FY 2026 revenue guidance is $7.650-$7.662 billion, implying 71% growth, with US commercial revenue above $3.224 billion (at least 120% growth), adjusted operating income of $4.440-$4.452 billion and adjusted free cash flow of $4.2-$4.4 billion. Q2 revenue guidance is $1.797-$1.801 billion. These are forecasts, not results.

Why Sovereign AI Is the Strategic Wager

Karp's public argument through the first half of 2026 is that frontier labs have oversold general-purpose models while absorbing enterprise data into their own systems. Speaking in early July, he described the typical enterprise experience as paying for tokens that produce no measurable value while handing over proprietary advantage.

If a defence ministry, intelligence agency or utility wants capable models but cannot send data to a commercial API, an open-weight model fine-tuned inside its own perimeter is the alternative. Palantir's pitch is that Ontology supplies the semantic and permissions layer that makes model output actionable rather than merely fluent, a framing chief technology officer Shyam Sankar summarised on the Q1 call as "the ontology is the body to the AI brains."

The opportunity is unquantified in Palantir's disclosures, and international revenue remains the softer part of the business: the United States supplied 79% of Q1 revenue, the United Kingdom 8% ($130.1 million) and the rest of the world 13%.

Where the Competition Sits

Palantir faces four categories of rival. Systems integrators including Booz Allen Hamilton and Accenture chase the same government budgets on a labour-hours model. Data-platform vendors such as Databricks and Snowflake compete on data and increasingly on AI tooling. Hyperscalers bundle model access, accreditation and compute. Large customers can build in-house.

Palantir's advantages are specific: accreditation for classified environments, forward-deployed engineers, and an installed base inside the US Army, where a consolidated enterprise service agreement worth up to $10 billion announced in mid-2025 replaced dozens of separate contracts. Its vulnerabilities are pricing and the improving quality of open-source alternatives, a sensitivity visible, when the shares fell about 6% partly on chatter about a free open-source tool with overlapping ambitions.

A Balance Sheet With Almost Nothing on It

Unlike the chipmakers and hyperscalers driving AI capital spending, Palantir does not build data centres. It held $2.29 billion in cash and equivalents plus $5.73 billion in marketable securities, roughly $8.0 billion, with no outstanding debt and a $500 million undrawn credit facility. Guided adjusted free cash flow of $4.2-$4.4 billion for 2026 would add materially to that.

The cost investors bear is dilution. Stock-based compensation was $201.6 million in Q1 2026, up 30% from $155.3 million a year earlier, meaningful, though far slower than the 85% revenue line. Weighted-average diluted shares rose about 1%, to 2.57 billion. Insider selling, including scheduled sales by Karp under trading plans, draws recurring criticism but reflects vesting mechanics as much as sentiment.

What Could Accelerate From Here

Conversion of the June sovereign-AI engine into disclosed agency contracts would be the cleanest validation. Expansion of the US Army enterprise agreement and related Pentagon work, broadened in March 2026, provides a large committed base, and Warp Speed, Palantir's manufacturing operating system, extends the commercial franchise into industrials. International defence remains underpenetrated: the UK Ministry of Defence signed a three-year, £240.6 million data-analytics extension effective recently under a defence and security exemption. Reports of a separate £750 million UK military AI award could not be reconciled with primary UK sources and should not be relied on.

The Risks Investors Should Price

Concentration comes first: US government work supplied 42% of Q1 revenue, exposing Palantir to appropriations cycles and political turnover. Second is reputational and regulatory friction. The UK's Office for Statistics Regulation intervened in July 2026 over how NHS England characterised benefits from the Palantir-built Federated Data Platform, and MPs have called for cancellation. Palantir's ICE work, including an approximately $30 million ImmigrationOS award, drew criticism from the Electronic Frontier Foundation, publicly rebutted by Palantir in April 2026, and questions from New York City pension funds. Third is model commoditisation. Fourth is expectation risk: at these multiples, deceleration alone could compress the stock even if results stay excellent.

What the Market Already Assumes

At the close of $131.53, market capitalisation was about $315.3 billion and enterprise value near $307.5 billion. On aggregated market data that equates to roughly 138 times trailing earnings, 77 times forward earnings, 60 times trailing sales and 59 times EV/revenue, among the highest multiples in large-cap US software, and figures that move daily.

Sell-side opinion is split. Across 32 analysts tracked by data aggregators in July 2026, the average 12-month target was $182.20, with a range from $70 to $255. Wedbush reiterated an Outperform rating and $230 target ; UBS carried a Buy rating and $200 target. These are individual firm forecasts, not company guidance.

Conclusion

Palantir is executing at a level almost no software company of its size has matched, and the Nvidia engine extends rather than transforms that position. But the launch is a product announcement, not a revenue event, and the market has already priced years of compounding into the shares. At roughly 60 times sales, the margin for disappointment is thin.

This article is for informational purposes only and does not constitute investment advice or a recommendation to buy, sell, or hold any security. Share prices can rise or fall, and past performance does not guarantee future results. Investors should conduct independent research and consult a licensed financial adviser before making investment decisions.