Palantir's 2025 results were genuinely strong: $4.5 billion in revenue, $1.6 billion in net income, $2.1 billion in free cash flow. The valuation leaves no room for anything to go wrong.
Palantir's 2025 financial results were genuinely impressive. Revenue grew 55% to $4.5 billion. Net income expanded from $493 million in 2024 to $1.6 billion in 2025. Free cash flow reached $2.1 billion with essentially zero capital expenditure. Operating margin jumped from 10.8% to 31.6% in a single year. The business is real, the growth is real, and the momentum is real.
At a $342 billion market cap, the valuation is also extraordinary. Price-to-earnings of 227x. Price-to-sales of 76x. EV-to-EBITDA of 245x. These are not slightly elevated multiples for a hyper-growth company. They are among the most extreme valuations ever assigned to a large-cap public company.
The risk analysis is not about whether Palantir is a good business. It is about what the 227x multiple requires to be correct, and what the distribution of outcomes looks like when the valuation is priced for perfection.
Palantir's narrative transformed in 2023 and 2024. For most of its public life, the company was viewed as a high-cost government contractor with a slow-growing commercial segment and a culture that made enterprise sales difficult. The stock spent extended periods trading at depressed levels despite its data integration capabilities.
The launch of AIP, the Artificial Intelligence Platform, changed the commercial narrative. AIP provides an integration layer that connects enterprise data to large language models using Palantir's ontology technology. The ontology, which creates a semantic model of how data objects relate to each other within an organization, gives AIP a differentiating angle compared to generic AI application builders.
AIP boot camps, multi-day intensive workshops where enterprise prospects experience rapid AI deployment use cases, drove customer acquisition. The format proved effective. Prospects who attended boot camps converted to paying customers at higher rates than traditional Palantir sales cycles. US commercial revenue began growing at 60% or more annually.
Revenue grew from $2.2 billion in 2023 to $2.9 billion in 2024 to $4.5 billion in 2025. The share count grew from 1.76 billion in 2020 to 2.57 billion in 2025, as Palantir issued shares to fund operations and compensate employees during its pre-profitability years. The dilution has slowed but not stopped: 2.57 billion shares outstanding today versus 2.53 billion a year ago.
The stock's inclusion in major indices and the retail trading community's enthusiasm for Palantir's narrative have created a valuation that the fundamental analysis cannot fully support at current levels. That gap between narrative and fundamental is the core risk.
TickerXray Report
Get the complete Palantir report with all 12 quantitative models, AI-generated investment thesis, and real-time data.
The improvement in Palantir's income statement from 2021 to 2025 is genuine and should not be minimized. The company went from $0.5 billion operating losses in 2021 to $1.4 billion operating income in 2025. Gross margins expanded from 78% in 2021 to 82.4% in 2025. The path to durable profitability that bears long doubted has been clearly established.
But the earnings quality deserves careful inspection. Stock-based compensation has run at approximately $0.7 billion annually in 2024 and 2025. Against reported net income of $1.6 billion, the $0.7 billion SBC represents 43.8% of net income. When analysts calculate cash earnings by adding SBC back to operating cash flow, the free cash flow of $2.1 billion looks impressive. But SBC is a real economic cost to shareholders through dilution, even if it does not consume cash.
The adjusted EPS of $0.63 TTM implies that new buyers at current prices are paying roughly $227 per dollar of earnings. If one strips SBC from the earnings calculation, the effective earnings yield is below 0.4%. That is a fraction of the risk-free rate.
Quarterly earnings beats have been substantial. Palantir beat estimates by 12.5% in Q2 2024, 11.1% in Q3 2024, 23.7% in Q4 2024, then 0% in Q1 2025, 14.3% in Q2 2025, 23.5% in Q3 2025, and 33.3% in Q4 2025. The beat pattern is consistent, which is genuinely a positive signal. But consensus estimates have been repeatedly adjusted upward, and the stock already trades at 227x the most recent consensus. Sustained beats at the current multiple do not guarantee multiple expansion.
At 227x earnings, Palantir must grow earnings at 35% to 40% annually for five to seven years to grow into its valuation on a reasonable basis. That is not continued success. It is sustained extraordinary performance for most of the current decade, starting from a base that is already large.
Government revenue, which still represents roughly half of total revenue, is not a 35% growth engine. US government revenue has grown as defense and intelligence spending increased, but procurement timelines are measured in years and budget cycles are political. International government revenue faces structural caps in budget-constrained allies.
The commercial segment, specifically US commercial, is the growth engine. Annual US commercial revenue growth above 60% has driven the narrative. But early-adopter cohorts are fundamentally different from mainstream enterprise customers. The AIP boot camp format works well when prospects are eager to experiment. The challenge is converting organizations that are skeptical, budget-constrained, or already committed to a competing AI platform.
If US commercial revenue growth decelerates from 60% to 30%, total company growth falls to roughly 20% to 25%. That is still fast growth by any normal standard. But the mathematics of valuation compression are unforgiving. A business growing at 25% with a clear platform advantage might fairly trade at 40x to 50x earnings. At 50x, the current 227x multiple compresses by 78%. On flat earnings, a multiple contraction from 227x to 50x produces an 78% drawdown. That is the structural risk embedded in the current price, even in a scenario where the business continues to perform well.
Palantir's share count trajectory tells a story that the earnings headline does not. Shares outstanding grew from 1.76 billion in 2020 to 2.57 billion in 2025, an increase of 46% over five years. Nearly half of that dilution came from SBC issued to employees and executives during the company's pre-profitability phase.
Dilution has slowed in recent years: 2.53 billion shares in 2024, 2.57 billion in 2025. That is approximately 1.6% annual dilution, lower than the historical pace. But at 2.57 billion shares, the fully diluted share count is substantially higher than many investors tracking per-share metrics realize.
Insider selling is worth noting with specificity. CEO Alex Karp sold 493,025 shares on February 20, 2026, at approximately $133.78 per share, generating roughly $66 million from a single transaction. CFO David Glazer sold 17,438 shares the same day. Systematic insider selling at elevated prices is a legal and often tax-driven activity. But when the CEO is selling nearly half a million shares in a single day and the stock is at 227x earnings, the signal is at least worth acknowledging.
The buy side is thin. Only one insider purchase appears in recent activity: a purchase by Ro Khanna, who is a US Congressman and not a company employee, reflecting his personal investment rather than insider conviction about the business outlook. The company's own executives and directors are net sellers at current prices.
Palantir's ontology-based data integration approach is genuinely differentiated, but differentiation in software erodes faster than in hardware. The competitive landscape has intensified meaningfully since AIP launched.
Microsoft, Salesforce, ServiceNow, and IBM are all building AI analytics and agent platforms that address the same enterprise use cases as AIP. Microsoft's Copilot for Azure and Power Platform integrations create an AI analytics layer for enterprises already on the Microsoft stack, which is most large enterprises. Salesforce Einstein and Agentforce address the CRM and customer analytics use case. ServiceNow's AI capabilities target IT operations and workflow automation.
None of these is a perfect substitute for Palantir's ontology technology. Palantir's advantage is the depth of its data modeling capabilities and its track record in complex operational environments. But enterprise AI budgets are finite. When a CISO or COO evaluates platforms, Palantir must justify a premium price against Microsoft or Salesforce, which the enterprise already pays for. That comparison is harder when the competition improves.
On the government side, Booz Allen Hamilton, Leidos, CACI, and Accenture Federal Services have all built or acquired AI analytics capabilities. The TITAN contract wins are real and important for Palantir's government revenue, but the competitive set for government AI contracts is broader than it was five years ago. Margin pressure from competition, even without significant market share loss, would compress Palantir's 31.6% operating margin.
Startups are also well-funded. Scale AI, Databricks, and vertical AI companies focused on healthcare, legal, or financial services are building specialized platforms that compete with Palantir in specific verticals. Palantir's breadth is an advantage, but focused competitors can outexecute on specific use cases.
The math of growing into a 227x multiple is demanding. Assume Palantir's stock is fairly valued when it reaches 35x earnings, a generous multiple for a mature software business. At that terminal multiple, Palantir needs to grow earnings from $1.6 billion to $9.8 billion to justify the current market cap. That is a 6x increase in earnings from a base that already reflects a step-change improvement in 2025.
If Palantir grows earnings at 35% annually, it reaches $9.8 billion in approximately nine years. In nine years, the current investor has earned the right to own a 35x earnings business at the price they paid. There is no incremental return from multiple expansion, only the earnings growth itself. Nine years of perfect execution to break even on valuation. That is the hurdle.
The analyst community has reached a clear conclusion. Fifteen holds, two sells, and four strong sells against only three strong buys and one buy. The consensus price target of $186.60 implies that the analyst consensus believes the stock is overvalued relative to fundamental value, even using optimistic estimates. This is unusual. Most large-cap technology companies have strong buy majorities. Palantir has a hold and sell majority among the analysts who cover it most closely.
Technology history offers relevant precedents. Cisco in 2000 was a genuinely great business at an extreme valuation. It took fifteen years to return to its peak price. Oracle traded at 80x earnings during its growth phase. It spent a decade below peak price before earnings caught up. The companies were not frauds. The valuations were simply pricing in futures that arrived on much longer timescales than implied.
Palantir's government revenue creates a political risk that most software companies do not face. The company's contracts with US Customs and Border Protection, Immigration and Customs Enforcement, and various intelligence agencies have attracted public controversy. Congressional scrutiny of AI surveillance tools is ongoing. A high-profile incident involving Palantir technology in a context that generates political backlash could disrupt government contract renewals in ways that are difficult to model.
Key person concentration is unusually high. Alex Karp's public profile, philosophical brand, and association with Palantir's identity is tightly coupled to how the company presents itself. His February 2026 sale of 493,025 shares, while legally routine, is a high-profile signal. His eventual departure, planned or otherwise, creates succession risk for a company whose culture and go-to-market approach is deeply personalized.
Geographic concentration is a structural vulnerability. US government and US commercial represent the vast majority of revenue growth. Palantir's international government business has grown slowly, partly because of the political sensitivities of selling intelligence-adjacent software to non-US governments and partly because international procurement cycles move slowly. If US government spending faces budget pressure or if the US commercial growth rate decelerates, there is limited geographic buffer.
Finally, the rising interest rate environment creates a direct headwind for high-multiple stocks. At 227x earnings, small changes in discount rates have large effects on intrinsic value calculations. When ten-year Treasuries yielded 1%, a 227x multiple for a high-growth business had theoretical justification in discounted cash flow models. At 4.3% yields, the same model produces a dramatically lower intrinsic value.
The near-term data points that will reveal whether the 227x multiple is sustainable or fragile are clear. First, Q1 2026 US commercial revenue growth. The consensus estimate implies continuation of 60%-plus growth. Any deceleration to 40% or below would represent a significant narrative disruption. Second, the Q1 2026 earnings beat. Estimates imply $0.28 of EPS. The Q4 2025 beat was 33%. If Q1 comes in near the estimate rather than well above it, the growth momentum story softens.
Third, government contract announcements. Palantir has been winning defense and intelligence contracts that represent medium-term revenue visibility. Additional TITAN program expansions, new NATO-related deployments, or expansion of civilian agency contracts would support the government revenue trajectory. Contract cancellations or budget cuts would be negative signals.
Fourth, SBC trajectory. If Palantir can continue growing earnings while holding or reducing SBC as a percentage of revenue, earnings quality improves. If SBC creeps back up as the company competes for AI talent, the gap between reported and economic earnings widens.
The bull case remains coherent: Palantir built the best enterprise AI deployment platform available, government and commercial customers are signing long-term contracts, and the operating leverage in the model will drive earnings growth that eventually justifies the multiple. That story needs four to six quarters of execution to begin compressing the multiple from 227x toward something less extreme.
Palantir built a genuinely differentiated platform and the 2025 results confirm that differentiation is translating into revenue and profit at scale. The five-year transformation from money-losing government contractor to $1.6 billion net income software business is real and should not be dismissed.
The valuation is pricing in a future that requires flawless execution for years at a pace that few large-cap technology companies have maintained. The analyst consensus of predominantly holds and sells, with a target price implying limited upside, reflects the judgment that even a well-executed bull case is already captured in the current price.
For investors who own Palantir at cost bases well below current levels, the decision is about risk tolerance and conviction in the commercial growth trajectory. For new investors evaluating entry at 227x earnings and 76x sales, the margin of safety is negligible, the downside from multiple compression is large, and the patience required to grow into the valuation is measured in years, not quarters.
Full forensic analysis of Palantir
+ 6 more models included
150,000+ stocks covered
Global coverage across 60+ exchanges. Every report includes all 12 quantitative models and AI analysis.
View plansEvery report runs 12 quantitative models and generates an AI investment thesis. From Piotroski scores to manipulation detection -- get the full picture in seconds.
12 forensic models
Piotroski, Altman, Beneish, DuPont & more
AI investment thesis
Synthesized outlook on every stock
Manipulation detection
Spot red flags before they hit the news
150,000+ tickers
Global coverage across 60+ exchanges
Expected return
Forward return projections for every stock
Real-time data
Live prices, insider trades, news sentiment
Free accounts get 1 report per month. Pro gets unlimited.
At a $342 billion market cap and 111x forward earnings, Palantir is trading on a commercial AI deployment thesis that has only partially shown up in the financials. Pulling apart the segment economics, the contract durations, and the customer cohort behaviour produces a more nuanced picture than the bull or bear case suggests.
The USDA contract and a fresh $300 million win look great in the headline. At 231x trailing earnings and 78x sales, the math tells a different story.
Revenue grew 56%, operating margin hit 41%, and FCF doubled to $2.1 billion. The stock has pulled back 30% from the high. The signal quality has not changed.