How Palantir Technologies Inc. Makes Money

What Palantir sells is not advertising or cloud storage or consumer devices — it sells the ability to see things that would otherwise remain invisible: patterns in logistics networks, anomalies in hospital medication dispensing, vulnerabilities in adversarial military formations. At its core, Palantir sells software platforms that ingest data from dozens, sometimes hundreds, of disparate sources — legacy government databases, ERP systems, IoT sensors, satellite feeds, financial transaction records — and create a unified, queryable layer on top of them. Palantir generates revenue through long-term software contracts, which the company refers to as committed contracts, supplemented by usage-based fees as platforms scale within customer organizations. Palantir's practice of embedding its own engineers at customer sites — a model it calls 'forward deployment' — creates a feedback loop between product development and real-world usage that is unmatched in enterprise software. If successful, this would transform Palantir from a direct-sales enterprise software vendor into a platform with network effects — a structural upgrade that could significantly expand both the total addressable market and the company's pricing power. He was reading deeply about the 9/11 Commission's preliminary findings and struck by the recurring theme in the intelligence community's post-mortems: the hijackers had left traces across multiple government databases, airline records, and financial systems — traces that, had they been connected, might have flagged the plot.

Long before Silicon Valley discovered the phrase 'data-driven decision-making,' a small team of engineers and mathematicians in Palo Alto was quietly building software that the CIA would fund through its venture arm and that American soldiers would use to track bomb-making networks in Iraq and Afghanistan. Investigators studying why American agencies missed the dots connecting the nineteen hijackers concluded, in part, that the problem wasn't a lack of data — it was the inability to connect it. That founding impulse — building software for the world's hardest data problems, primarily in service of government and military clients — shaped Palantir into something genuinely unlike any other technology company in American history. The launch of the Artificial Intelligence Platform (AIP) in April 2023 — and the aggressive 'boot camp' sales strategy that followed — turbocharged Palantir's commercial growth in the United States. In the age of AI, that capability has become one of the most commercially and strategically valuable things in the world, and Palantir, after twenty years of building it in the dark, is finally getting the recognition — and the stock price — to match. With a market capitalization exceeding 170 billion dollars as of early 2025 and a forward price-to-sales ratio that reflects extreme investor optimism about AI monetization, Palantir remains one of the most debated, admired, and controversial technology companies in the United States. US Commercial revenue in FY2024 reached approximately 702 million dollars, growing 54 percent year-over-year in Q4 2024 alone. This segment has been the fastest-growing in Palantir's recent history, driven almost entirely by the adoption of AIP. Palantir's AIP boot camps — intensive three-to-five day working sessions where Palantir engineers embed with a client's team to build a working AI application — have proven remarkably effective at converting prospects into paying customers. International Commercial revenue was approximately 406 million dollars, a segment that has grown more slowly than US Commercial due to differing data privacy regulations (particularly GDPR in Europe) and longer sales cycles. The launch of AIP in April 2023 represented a genuine architectural inflection point for Palantir's business model. Palantir disrupted this model by offering intensive, hands-on workshops where a prospective customer's own team builds a real AI application on AIP within three to five days, using their own data, solving their own problem. The key competitive tension is between Palantir's software-first, commercial-off-the-shelf approach and the primes' tendency to build custom, government-specific solutions that require armies of human contractors to maintain. Snowflake, a public company with FY2024 revenue of approximately 3.6 billion dollars, dominates cloud data warehousing and has expanded into AI applications through its Cortex platform. Palantir's answer is the Ontology and the governance layer: that generic AI APIs require enormous internal engineering investment to operationalize safely for enterprise and government use, and that Palantir's platform absorbs that complexity so customers don't have to. Palantir's financial story from 2020 to 2024 is one of gradual commercial diversification accelerating sharply into something resembling a genuine growth inflection. Revenue growth was steady but not spectacular through 2021 and 2022, with FY2021 revenue reaching 1.542 billion dollars (41 percent growth) and FY2022 revenue reaching 1.906 billion dollars (24 percent growth). The 2022 period was difficult for Palantir shareholders: rising interest rates crushed growth stock multiples broadly, and Palantir's shares fell more than 70 percent from their early 2021 highs, touching a low of approximately 6 dollars per share in December 2022. FY2023 marked the beginning of the AIP-driven acceleration, with revenue of approximately 2.228 billion dollars (17 percent growth) and the first quarters of GAAP profitability. US commercial revenue in Q4 2024 grew 54 percent year-over-year to 214 million dollars — a growth rate that, if sustained, would double that segment within eighteen months. Federal contract renewals are subject to political cycles, budget appropriations, and shifting procurement priorities in ways that commercial SaaS contracts are not. Analysts at major investment banks have noted that at current multiples, Palantir must sustain revenue growth above 25 to 30 percent annually for the better part of a decade to justify the stock price — a bar that even exceptional software companies rarely clear. While Palantir has made progress reducing SBC as a percentage of revenue, the culture of equity compensation is deeply embedded in the company's talent strategy, and meaningfully reducing it risks impairing the company's ability to attract and retain the specialized talent — cleared data scientists, AI engineers, former intelligence officers — that underpins its value proposition. Palantir's relationships with Immigration and Customs Enforcement (ICE), predictive policing programs in cities like New Orleans and Los Angeles, and various surveillance-focused government clients have generated sustained activist opposition, academic criticism, and media scrutiny. The tension between building tools for national security and enabling what critics characterize as mass surveillance is unlikely to diminish as Palantir expands its government footprint, and it creates ongoing reputational risk with potential commercial clients in industries sensitive to public perception, such as consumer healthcare and financial services. Building an Ontology for a complex organization — say, the US Army's logistics network, which spans hundreds of bases, thousands of vehicles, and millions of inventory items — requires years of close collaboration between Palantir engineers and the client. Palantir's growth strategy for 2025 and beyond rests on four pillars that are interconnected and mutually reinforcing. First, US commercial customer acquisition through AIP boot camps remains the primary near-term growth driver. The company has invested heavily in scaling its boot camp capacity, training more Palantir engineers to run them and expanding the boot camp curriculum to cover more vertical use cases — from hospital operations to oil and gas asset management to financial risk analytics. The company has been more selective about international commercial investments, focusing on markets with favorable data sovereignty regulations and existing enterprise relationships. Palantir has begun releasing developer tools that allow third-party software companies to build applications on top of the Ontology, creating a potential marketplace dynamic similar to Salesforce's AppExchange. Palantir's management issued FY2025 revenue guidance of approximately 3.74 to 3.76 billion dollars, representing guidance-implied growth of approximately 31 percent — an acceleration from FY2024's 29 percent. The most immediate growth driver is the continued expansion of AIP boot camps and the conversion of boot camp attendees into paying customers. The defense AI opportunity is also expanding rapidly. The US military's Maven Smart System program, to which Palantir is the primary software contributor, is being expanded to additional commands and allied militaries. Thiel's insight was that this problem had a solution that the government, with its procurement timelines and institutional inertia, could never build itself. Karp himself has argued that his training in critical theory gave him an ability to see Palantir's work in its full social and political context — to think about the implications of building surveillance infrastructure, not just the technical challenges of building it. The company's earliest funding came not just from Thiel but from In-Q-Tel, the CIA's nonprofit venture investment arm established in 1999 to help the intelligence community access commercial technology. In-Q-Tel's investment in Palantir — variously reported at between 2 million and 14 million dollars, with the precise figure remaining undisclosed — was an unusual arrangement that gave Palantir access to operational requirements from intelligence agencies, effectively turning real national security problems into product specifications. The arrangement also gave the CIA insight into Palantir's product roadmap and created a customer relationship that would generate revenue for years. Critics have noted that the In-Q-Tel structure means Palantir was, from its earliest days, building software explicitly for government surveillance and intelligence applications with government guidance on what those tools should do — which raises questions about the company's claimed status as a neutral software vendor rather than an extension of the national security apparatus.