Palantir has spent years dividing Wall Street into two camps. To believers, it is becoming the operating system for artificial intelligence inside governments, factories and some of the world’s most complex organizations. To skeptics, it is a highly valued software company wrapped in defense contracts, theatrical shareholder letters and an unusually devoted investor following.
Its latest results gave the believers considerably more ammunition.
Palantir’s growth is no longer merely holding up against difficult comparisons. It is accelerating at a scale rarely seen in enterprise software, while profitability and cash generation are expanding alongside revenue. That combination helps explain why investors increasingly talk about Palantir as an AI winner rather than a niche government contractor.
The difficult question is no longer whether the business has momentum. It clearly does. The question is how much future success has already been priced into the stock — and what happens when extraordinary growth eventually becomes merely excellent.
Another quarter, another surprise
Palantir’s second-quarter 2026 numbers were not a routine earnings beat. They represented another sharp upward reset of what investors thought the company could deliver.
Revenue reached $1.94 billion, up 93% year over year, compared with the roughly $1.8 billion expected by analysts. Adjusted earnings came in at $0.41 per share, ahead of the $0.35 consensus estimate.
The underlying operating figures were even more striking. U.S. revenue increased 115% to $1.57 billion, with U.S. commercial revenue rising 149% to $764 million and U.S. government revenue climbing 90% to $809 million.
Palantir also closed 220 contracts worth at least $1 million during the quarter. Seventy-three were valued at $10 million or more, while total contract value reached $3.37 billion. Its U.S. commercial remaining deal value — a measure of contracted business that has not yet been recognized as revenue — rose 124% to $6.24 billion.
Most software companies would be pleased to post that level of growth while burning cash to capture market share. Palantir is doing it with margins that look more like those of a mature platform monopoly.
The company reported a 47% GAAP operating margin and a 62% adjusted operating margin. Adjusted free cash flow reached $1.22 billion, equivalent to 63% of quarterly revenue. Its so-called Rule of 40 score, which combines growth and profitability, hit 155%.
Palantir subsequently increased its full-year revenue forecast to between $8.15 billion and $8.158 billion, up from its previous range of $7.65 billion to $7.662 billion. It now expects U.S. commercial revenue to exceed $3.424 billion, representing growth of at least 134%.
The latest figures and guidance are available in Palantir’s second-quarter earnings release filed with the SEC.
This is the central reason Palantir keeps beating expectations: analysts are raising their forecasts, but the company’s growth is accelerating even faster.
Government AI is finally going mainstream
Palantir’s government business was once treated as both its main advantage and its biggest limitation.
The advantage was obvious. Government relationships are difficult to win, security requirements create high barriers to entry, and successful deployments can lead to long-running contracts. The limitation was that public-sector procurement is slow, politically sensitive and often dependent on a relatively small number of large customers.
That picture is changing.
Defense departments are no longer experimenting with AI as a distant research project. Modern military operations increasingly depend on integrating satellite imagery, sensor data, logistics, targeting systems and battlefield communications in real time. Palantir’s software is designed to organize that fragmented information and turn it into operational decisions.
The company is therefore positioned closer to the command layer of AI than to the chatbot layer consumers normally associate with the technology.
Geopolitical instability is accelerating that demand. Governments are investing in autonomous systems, intelligence platforms, supply-chain resilience and software-defined defense infrastructure. Palantir’s involvement in projects connected with battlefield systems and the proposed U.S. Golden Dome missile-defense initiative gives it exposure to spending categories that could remain strategically important for years.
But defense is no longer the entire explanation for the company’s performance.
U.S. commercial revenue is now growing considerably faster than government revenue. Palantir has taken technologies originally built for intelligence agencies and adapted them for manufacturers, hospitals, energy companies, financial institutions and other businesses with complicated data environments.
Its Artificial Intelligence Platform, or AIP, allows organizations to connect AI models with their internal data, software and decision-making processes. Crucially, Palantir is not simply selling access to a model. It is selling the infrastructure required to make models usable inside organizations where data permissions, security, auditability and operational reliability matter.
The company’s intensive “bootcamp” sales strategy has also helped shorten adoption cycles. Instead of asking customers to commit to a lengthy transformation project before seeing results, Palantir places engineers alongside a company’s staff and attempts to build a working use case within days.
That approach is more labor-intensive than a conventional self-service software sale. It can, however, demonstrate value quickly and turn an experimental AI budget into a much larger deployment.
Palantir’s recent numbers suggest that this conversion process is working.
The valuation problem
A great company does not automatically make a great stock at every price.
Following its latest report, Palantir remained one of the most highly valued large software companies in public markets. Estimates vary with the share price and methodology, but the company was trading at roughly 45 times trailing revenue and more than 80 times trailing free cash flow in early August 2026.
Those figures are difficult to justify with ordinary software growth. Palantir must continue delivering something close to extraordinary performance.
The bullish argument is that conventional valuation comparisons fail to capture the company’s unusual combination of revenue acceleration, enormous margins and exposure to two powerful spending cycles: enterprise AI adoption and modern defense.
That argument has merit. A business growing near triple-digit rates while producing substantial free cash flow deserves a higher multiple than a mature software vendor expanding at 10% or 15% annually.
The problem is that a valuation this rich does not merely assume Palantir will remain successful. It assumes the company will become dramatically larger while preserving exceptional margins and avoiding meaningful competitive pressure.
Consider the mathematics. Even if the share price stopped rising entirely, Palantir’s revenue would need to multiply before its sales multiple fell into the range normally associated with premium software companies. Any further increase in the stock raises that performance threshold again.
High valuations also make earnings beats less informative than they appear. Palantir can exceed quarterly expectations and still disappoint investors if its guidance implies that growth is beginning to normalize. When the market has priced in near-perfection, “better than expected” may not be better than the expectations embedded in the stock.
This does not mean Palantir is inevitably overvalued. It means investors are paying today for several years of execution that have not happened yet.
The AI gold rush’s quiet winner
Palantir occupies a different part of the AI economy from Nvidia, OpenAI or Anthropic.
Nvidia sells much of the computing infrastructure used to train and run AI models. The leading model developers build the foundation models that generate text, images, code and other outputs. Cloud providers rent the infrastructure required to operate them.
Palantir’s opportunity sits further downstream.
Businesses may have access to powerful models, but those models are not automatically connected to inventory systems, factory schedules, classified databases or hospital workflows. They also do not inherently understand who has permission to access specific information or which actions require human approval.
Palantir attempts to solve that deployment problem.
Its platforms create a digital representation of an organization’s operations, sometimes described as an ontology. Data from different systems can then be connected to real-world assets, employees, processes and decisions. AI models operate within that governed environment instead of floating above it as isolated tools.
This gives Palantir an attractive position in the AI stack. It does not have to fund the increasingly expensive race to train the world’s largest foundation model. Nor does it manufacture chips, build fabrication plants or carry the same level of hardware-cycle risk as semiconductor companies.
Instead, it benefits when models become more capable and when customers become more eager to deploy them.
Palantir can therefore win even in a world where foundation models become cheaper and increasingly interchangeable. In fact, falling model costs could improve its value proposition by making AI-powered workflows economical across more industries.
Its real product is not raw intelligence. It is the controlled application of intelligence inside organizations that cannot afford unreliable answers or uncontrolled access to sensitive data.
That distinction is one reason Palantir has emerged as a quieter winner of the AI gold rush. It sells the picks and shovels required after customers discover that owning an AI model is not the same thing as transforming a business.
What could break the Palantir story?
Palantir’s current momentum is formidable, but the investment case still contains several obvious failure points.
The first is growth deceleration. Revenue cannot expand near 90% indefinitely as the comparison base becomes larger. The market may tolerate a gradual slowdown, especially if profits continue rising. A sudden drop in commercial bookings or weaker-than-expected guidance would be far more damaging.
The second is customer concentration and procurement risk. Government contracts can be large and durable, but they can also be delayed, challenged or redirected for political reasons. Palantir has already faced resistance in parts of Europe as governments debate technological sovereignty and dependence on American platforms.
The third is competition. Cloud providers, consultancies and enterprise software companies all want to become the control layer for corporate AI. Customers may also build more capabilities internally as models and development tools improve.
Palantir’s advantage depends on remaining faster, safer and more useful than those alternatives.
The fourth is execution at scale. Its hands-on deployment model has helped win customers, but the company must continue expanding without turning into a consultancy whose growth depends on adding engineers at the same rate as revenue.
The fifth is political and reputational exposure. Palantir’s work with defense, intelligence and immigration agencies regularly attracts scrutiny. Some customers or employees may avoid the company because of those relationships, while regulatory disputes could complicate future contracts.
Finally, there is the valuation itself.
Even without a collapse in the underlying business, the stock could fall sharply if investors decide that a lower multiple is appropriate. Interest rates, market sentiment or a broad retreat from AI stocks could compress Palantir’s valuation long before its revenue or profits deteriorate.
Palantir does not need to fail for shareholders to lose money. It only needs to perform less spectacularly than the current price demands.
The company’s results show that the AI momentum is not imaginary. Commercial adoption is accelerating, government demand is expanding, margins are exceptional and contract activity points to a substantial pipeline.
But the comparison with Nvidia should be treated carefully. Nvidia became the defining stock of the first phase of the generative AI boom because nearly every major AI developer needed its hardware. Palantir is attempting to become essential during the next phase: deploying AI inside real institutions.
That opportunity could be enormous. The stock price already assumes that Palantir will capture a meaningful part of it.
Palantir AI Stock: Cyberpunk Boom or Bubble?
Palantir stock: key questions
Why does Palantir keep beating earnings expectations?
Demand is growing faster than analysts have anticipated, particularly among U.S. commercial and government customers. Strong contract activity and rapid AIP deployments are also allowing Palantir to repeatedly increase its guidance.
Is Palantir only a defense contractor?
No. Government work remains central to the business, but U.S. commercial revenue is currently growing faster. Palantir sells software to companies across manufacturing, healthcare, energy, finance and other sectors.
How is Palantir different from Nvidia?
Nvidia primarily supplies the chips and computing systems behind AI. Palantir provides software that connects AI models with an organization’s data, permissions and operational workflows.
Why is Palantir’s valuation considered so high?
Its market value represents many times its annual revenue and free cash flow. That pricing assumes years of rapid growth, expanding profits and continued leadership in enterprise and government AI.
What is the biggest risk for Palantir investors?
A slowdown in growth could trigger a major valuation reset. Competition, government procurement changes, political controversy and weaker commercial demand could also damage investor sentiment.
Is Palantir stock financial advice?
No. This article is for informational purposes only and does not constitute financial or investment advice. Investors should assess their objectives, time horizon and tolerance for volatility before making any decision.