$PLTR Stock Bull Case: Palantir Is Becoming AI Infrastructure

Palantir just delivered the kind of quarter that forces investors to update their assumptions.

Revenue reached $1.935 billion, up 93% from a year earlier. U.S. commercial revenue grew 149%. U.S. government revenue grew 90%. GAAP operating margin hit 47%, while adjusted operating margin reached 62%.

Most software companies can produce explosive growth or serious profitability. Palantir is now producing both at the same time.

That is the center of my bull case.

I do not see Palantir as another company selling an AI assistant. I see it becoming infrastructure for organizations that need artificial intelligence to make decisions, coordinate people, protect sensitive data, and take action inside the real world.

That distinction matters.

A chatbot can summarize a document. Palantir wants to help a hospital allocate beds, help a manufacturer respond to a supply disruption, help an energy company manage equipment, and help a military organization coordinate missions.

The opportunity gets much larger when AI moves from answering questions to running workflows.

The numbers suggest that customers are making that move faster than many investors expected. The valuation, however, leaves very little room for mistakes. My view is bullish on the business and disciplined on the stock.

Let me start with the part bulls often avoid.

A great company can still become a bad investment when the stock price assumes years of nearly perfect execution.

Palantir is priced as an exceptional company because it is producing exceptional results. That does not make valuation irrelevant. It makes valuation more dangerous.

Investors are already expecting rapid growth, expanding margins, major commercial adoption, durable government demand, and continued leadership in operational AI. If any part of that story weakens, the stock could fall sharply even while the company remains profitable and continues growing.

That is the strongest bear argument, and it deserves respect.

The correct question is not whether Palantir looks expensive compared with an average software company. It clearly does. The question is whether Palantir can grow into that premium by becoming a foundational operating layer for AI.

I think the evidence is moving in that direction. I also think anyone researching the stock should separate confidence in the company from comfort with the current price.

Use the MentorSurge Market Dashboard to check the current quote and market trend before making any conclusion. Price changes faster than the business thesis.

Palantir does not manufacture chips, build data centers, or train every major language model. Calling it AI infrastructure does not mean it competes directly with Nvidia or replaces the cloud providers.

It occupies a different layer.

The expensive problem facing large organizations is not merely gaining access to a capable model. Models are increasingly available from several vendors. The harder problem is connecting those models to fragmented data, company rules, employee permissions, physical operations, and accountable decisions.

That is where Palantir is positioning itself.

An enterprise may have customer records in one system, inventory data in another, factory information somewhere else, and security rules spread across the organization. Giving an AI model access to everything without control would be reckless. Keeping the model isolated from operating data would make it far less useful.

Palantir builds the layer between those two extremes.

Its software connects data, logic, actions, and security. The goal is not simply to generate an answer. The goal is to let people and AI agents work with a shared model of the organization, while controlling who can see information and what actions can be taken.

That is a much deeper product than a chat window.

Palantir becomes easier to understand when I separate the platform into three parts.

Foundry connects data from different systems and turns it into something people can use.

At the center is Palantir’s Ontology. The Ontology represents the important objects, relationships, logic, actions, and security rules inside an organization.

For an airline, those objects could include aircraft, crews, gates, maintenance schedules, and passenger connections. For a manufacturer, they could include factories, machines, suppliers, parts, orders, and delivery dates.

The value is not just collecting data. The value is connecting data to decisions and actions.

Once those relationships are built into daily workflows, replacing the platform becomes more difficult. The customer would not simply be changing a reporting tool. It could be rebuilding the operational language used across the organization.

AIP brings language models and AI agents into that controlled operating environment.

Organizations can connect approved models to their data, build applications and agents, evaluate their behavior, and govern what those systems are allowed to do.

Palantir does not need one model provider to win every race. Its documentation says AIP supports models from several leading providers. That gives customers flexibility while Palantir remains the layer connecting models to operations.

This is a powerful position.

The model may change. The cloud provider may change. The workflow, permissions, institutional knowledge, and decision structure can remain inside Palantir.

Apollo handles software deployment and continuous updates across cloud environments, customer infrastructure, and edge locations.

That matters for customers that cannot place every workload inside one public cloud. Governments, defense organizations, hospitals, manufacturers, and other regulated institutions may need software to operate across unusual or highly controlled environments.

According to Palantir’s architecture documentation, Apollo manages the infrastructure supporting Foundry and AIP and coordinates thousands of upgrades across services and assets.

Together, AIP, Foundry, and Apollo look less like three separate products and more like an operating system for modern organizations.

The most important number in the latest quarter was U.S. commercial revenue.

It reached $764 million, rising 149% from the prior year and 28% from the previous quarter.

That level of growth changes the old Palantir narrative.

For years, critics viewed the company mainly as a government contractor with long sales cycles and a limited commercial future. The latest results show a commercial business growing faster than the government segment and becoming a major engine of the company.

Contract activity supports that argument.

Palantir closed 220 deals worth at least $1 million during the quarter. Of those, 98 were worth at least $5 million and 73 were worth at least $10 million.

U.S. commercial total contract value reached $2.132 billion, up 153% from a year earlier. U.S. commercial remaining deal value reached $6.238 billion, up 124% from a year earlier and 27% from the prior quarter.

The customer base expanded too. Palantir reported 1,049 customers for the trailing twelve month period, compared with 849 a year earlier.

This looks like more than a handful of experimental AI projects. It suggests that customers are moving from pilots into larger operational deployments.

Still, deal value is not the same as recognized revenue. Some contracts include options, cancellation rights, or other conditions. I treat contract metrics as evidence of demand, not as guaranteed future sales.

U.S. government revenue reached $809 million in the quarter, up 90% from the prior year and 18% from the previous quarter.

That is not a sleepy legacy segment.

Government adoption gives Palantir revenue, credibility, and experience in environments where accuracy, security, and reliability matter. Work completed in demanding government settings can strengthen products that later reach commercial customers.

The relationship also works in reverse. Improvements created for commercial customers can strengthen the platform used by governments.

This creates a useful feedback loop.

The risk is that government revenue can be affected by budgets, procurement delays, policy changes, contract options, and shifts in political priorities. The company also warns that many government contracts can be terminated for convenience.

The government segment should therefore be treated as durable but not automatic.

For the first six months of 2026, government customers produced 52% of total revenue and commercial customers produced 48%. That balance is healthier than a business dependent on only one customer type.

Rapid revenue growth is more valuable when it produces real operating leverage.

Palantir reported $912 million of GAAP operating income in Q2, equal to a 47% margin. GAAP net income attributable to common stockholders reached $1.062 billion.

Cash generation was also strong. Cash from operations reached $1.216 billion, while adjusted free cash flow reached $1.220 billion.

The company ended the quarter with $9.2 billion in cash, cash equivalents, and short term U.S. Treasury securities. Its filing also showed no outstanding balance under its $500 million revolving credit facility.

That balance sheet gives Palantir room to invest, hire, expand infrastructure, and absorb economic volatility without depending on outside financing.

Management highlighted a Rule of 40 score of 155%, calculated from the company’s 93% revenue growth and 62% adjusted operating margin. The measure is not GAAP, and it should not replace the financial statements. It does show how unusual the current combination of growth and profitability has become.

The 85% GAAP gross margin is another important signal. It suggests that even as cloud hosting expenses rise, the underlying software model retains strong economics.

This is what separates the current Palantir story from the version investors debated a few years ago. The company is not asking the market to wait indefinitely for operating leverage. It is showing that leverage now.

Management guided Q3 revenue to between $2.160 billion and $2.164 billion.

For the full year, it raised expected revenue to between $8.150 billion and $8.158 billion. It expects U.S. commercial revenue to exceed $3.424 billion, representing growth of at least 134%.

Adjusted operating income guidance rose to between $4.889 billion and $4.897 billion. Adjusted free cash flow guidance moved to between $4.5 billion and $4.7 billion.

Palantir also continues to expect GAAP operating income and net income in every quarter of the year.

Guidance can change, and management still has to execute. But these numbers indicate that the second quarter was not presented as a temporary spike. The company expects the momentum to continue.

That matters because hypergrowth can look impressive for one quarter and vanish when a large contract shifts timing. Palantir is telling investors that the growth engine has enough depth to carry through the rest of the year.

The next earnings report will test that confidence.

The strongest part of Palantir’s moat is not a single model or feature.

It is the combination of technology, deployment experience, security, and customer integration.

Palantir engineers work close to customer problems. That field experience feeds product development. The resulting platform is then reused across industries and use cases.

Once a customer has built an Ontology, connected critical systems, defined permissions, trained employees, and embedded applications into daily decisions, switching becomes costly.

This is not only technical lock in. It is organizational lock in.

The platform begins to contain a map of how the institution operates. It captures relationships between data, decisions, people, and actions. Competitors can offer capable models or data tools, but recreating that operating context takes time.

That is why I think comparisons with ordinary software vendors miss the point. Palantir is trying to become the system through which organizations understand and operate their world.

For the hardware side of the AI buildout, read the MentorSurge Nvidia bull case. For the cloud and distribution layer, read the Amazon three engine bull case.

A serious bull case needs clear failure points.

Palantir could continue growing while the stock falls because investors decide to pay a lower multiple. The more optimism included in the price, the more violent that reset can become.

Current growth is extraordinary. It will eventually slow. The key question is whether it settles at a level that still supports the valuation. A sudden drop in U.S. commercial growth would weaken the central thesis.

The company explicitly warns that it may not realize the full value of customer contracts. Options may not be exercised, contracts may be renegotiated, and some government agreements may be terminated.

RDV is useful, but it is not a bank account.

Stock based compensation reached $265 million in the quarter, up 66% from a year earlier. For the first half, it reached $467 million, up 48%.

Basic weighted average shares rose from about 2.365 billion to 2.400 billion. That increase was modest compared with revenue growth, but dilution still matters because each share represents a claim on the business.

The company also had $916 million of unrecognized compensation expense tied to outstanding restricted stock units at quarter end. I want revenue and cash flow per share to improve, not only company totals.

No single customer represented more than 10% of quarterly revenue, which is encouraging. However, one customer represented 27% of accounts receivable at quarter end.

The top twenty customers produced average trailing twelve month revenue of $124 million each, up 67% from a year earlier. That shows powerful expansion, but it also means large relationships matter.

U.S. customers generated 80% of revenue during the first half of 2026. That concentration has fueled growth, but it leaves international expansion as both an opportunity and a question.

Palantir amended a cloud services agreement in March 2026 that requires at least $5.6 billion of spending over ten contract years through February 2036.

That commitment can support scale if demand continues growing. It becomes a burden if usage falls below expectations. The filing says annual minimum commitments range from $268 million to $979 million.

Palantir’s share structure gives its founders unusual influence.

The company explains that its Class F structure can give the founders effective control over major matters submitted to stockholders. Public investors receive the economics of ownership without the same level of voting influence.

Some investors will accept that arrangement because they trust management’s long term vision. Others will see it as a governance discount. Both views are reasonable.

Palantir competes with cloud providers, data platforms, software vendors, internal engineering teams, and new AI companies.

Its current execution is impressive, but the market will not stand still. The bull case requires Palantir to keep proving that its integrated platform produces better operational results than cheaper or simpler alternatives.

For another skeptical view of AI valuation, read Why Michael Burry Has Got the AI Trade Wrong.

Here is what I would track every quarter.

That checklist keeps the thesis connected to business performance instead of social media excitement.

Palantir is becoming one of the clearest examples of AI moving from experimentation into operations.

The company is not merely selling access to a model. It is building the controlled environment where data, models, people, and actions can work together.

Q2 provided strong evidence. Revenue grew 93%. U.S. commercial revenue grew 149%. U.S. government revenue grew 90%. GAAP operating margin reached 47%. Cash generation was substantial, and management raised full year guidance.

Those are not hype metrics. They are business results.

The valuation remains the pressure point. Palantir has earned a premium, but no company earns an unlimited one. A bullish thesis should survive without assuming that investors will always pay a higher multiple.

My view is simple. Palantir’s business is getting stronger, its platform is becoming more important, and its role in operational AI is expanding. The stock still requires discipline because the market already understands much of that promise.

That tension is exactly what makes $PLTR worth studying.

Follow @mentor_surge on Instagram for the visual breakdown and the next research update.

This article is education and research only. It is not personalized investment advice, a recommendation to buy or sell $PLTR, or a promise of performance. Investing involves risk, including loss of principal.