Palantir: The Most Controversial Tech Company in the World

The Company Nobody Expected to Become an AI Story

Twenty-three years ago, a small group of Silicon Valley engineers took seed money from the CIA's own venture arm and started building software to help American intelligence agencies make sense of terrorist networks. Today that company is a $309 billion publicly traded AI bet, one of the most heavily discussed names on Wall Street, and arguably the single most divisive stock in the market. Both of those sentences describe the same company: Palantir Technologies.

Palantir occupies a strange position in the market. It is, simultaneously, one of the best-performing large-cap stocks of the past four years, a genuine and differentiated AI infrastructure business, and the subject of sustained, serious criticism from civil liberties organizations over its role in government surveillance and immigration enforcement. Most equity research treats these as separate stories — the growth story for the bulls, the ethics story for the critics. They are not separate. They are the same company, and understanding Palantir means holding both at once. This report covers what Palantir actually builds, how the business has performed, what it currently costs, and precisely what the controversies are — using the company's own numbers and named contracts, not vague characterizations.

What Palantir Actually Is

Palantir was founded in 2003 by Peter Thiel — the PayPal co-founder and early Facebook investor — together with Alex Karp, who has served as CEO ever since, and a small founding team. The name comes from the palantíri, the all-seeing seeing-stones in J.R.R. Tolkien's The Lord of the Rings — a naming choice that, in hindsight, reads less like a coincidence and more like a mission statement.

The company's first institutional backer was In-Q-Tel, the CIA's own venture capital arm, created specifically to fund technology useful to US intelligence. Palantir's first customers were not enterprises — they were intelligence agencies: the NSA, the CIA, the FBI. For roughly its first decade, Palantir operated in near-total secrecy, avoiding press coverage and public disclosure as a matter of policy. It built software to help government analysts pull signal out of enormous, fragmented datasets — tracking terrorist financing networks, mapping supply chains, coordinating battlefield logistics.

That origin story is not incidental colour. It is the reason Palantir's software works the way it does, the reason government remains its largest and stickiest customer base, and the reason the company draws a level of scrutiny that a typical enterprise-software vendor never would. A company that built its first product for the CIA does not get to be evaluated purely as a SaaS business — and to its credit, Palantir has never really tried to hide that history.

Palantir's public register is also unusual for a large-cap tech company, and it matters for how the stock trades. CEO Alex Karp writes openly about the company's alignment with Western governments and its willingness to work on defence and security problems that many Silicon Valley firms actively avoid. Co-founder Peter Thiel has built one of the most influential political and investment networks in American technology, extending well beyond Palantir into venture capital and politics. Most enterprise software companies try to stay studiously boring and apolitical. Palantir has never done that, and the market has, over time, come to price the company partly on that identity — for both bulls and critics.

The Real Technical Moat: Ontology

It is worth being precise about why Palantir's software is hard to replace, because "data integration" undersells it. Most data platforms move information between systems and leave it at that — a pipe, not a picture. Palantir's core technical concept is the ontology: a persistent, live digital model of an organisation's real-world objects — a specific aircraft, a specific hospital patient, a specific shipment, a specific soldier's unit — and the relationships between them, built directly on top of the underlying fragmented databases without requiring them to be re-architected first.

Once that ontology exists, an analyst or an AI agent isn't just querying rows in a table; they're reasoning about the actual entity the data represents, with the relationships already mapped. That is the mechanism behind Palantir's claim that once it's embedded, it becomes close to irreplaceable: ripping Palantir out doesn't just mean migrating a database, it means rebuilding the institution's live operational model of itself from scratch. That is a materially higher switching cost than a typical SaaS contract, and it is the real, defensible core of the bull case — not just brand recognition or government relationships.

The Two Products: Gotham and Foundry

Palantir's software is genuinely different from a typical business-intelligence tool, and understanding the difference matters for the investment case. Most enterprise software is built to operate on data that already lives in one clean, structured system. Palantir's core insight was that real institutions — governments, hospitals, airlines, militaries — never have clean, unified data. They have dozens of incompatible systems, spreadsheets, sensor feeds, and legacy databases that were never designed to talk to each other. Palantir's platforms exist to stitch that mess into one operational picture that a human can actually act on.

There are two products. Gotham is built for defence and intelligence customers — the platform tracks networks, models battlefield logistics, and flags financial fraud patterns across datasets that were never meant to be combined. Foundry is the commercial equivalent, sold to enterprises that face the same fragmented-data problem without the classified-environment requirements.

The customer list is genuinely broad for a company with this origin story: the US Army, the UK's National Health Service, Delta Air Lines, Ferrari, and Airbus are all named Palantir customers. When COVID-19 hit, the UK government used Palantir's software to help coordinate its national vaccine rollout — logistics, supply, and prioritisation across a system with no single owner. Palantir's software has also reportedly been used in targeting decisions during the Ukraine conflict, a claim the company has not denied and that sits at the center of ongoing debate about the role of commercial software vendors in modern warfare.

The AI Pivot: AIP

For most of its history, Palantir was a data-integration company, not an AI company in the modern sense. That changed in 2023, with the launch of AIP — the Artificial Intelligence Platform.

AIP's core idea addresses a problem that has become central to enterprise AI adoption: how do you get the benefits of large language models without sending your most sensitive data to an external provider like OpenAI or Google? For most companies, that trade-off is uncomfortable but tolerable. For a defence contractor, an intelligence agency, or a bank handling regulated financial data, it can be disqualifying. AIP lets an organisation run large language models directly on top of its own private data, inside its own security perimeter — no data ever has to leave the building.

That is, in plain terms, the entire commercial pitch: the power of modern AI, without the security exposure. For exactly the kind of customer Palantir has always served — governments, defence primes, regulated financial institutions — that combination is close to unique. AIP has become the growth engine behind the company's re-rating from a slow-growth government contractor into one of the market's highest-profile AI infrastructure names.

The go-to-market motion behind AIP is also unusual, and it's worth understanding because it explains why commercial adoption accelerated as fast as it did. Rather than a traditional months-long enterprise sales cycle, Palantir runs what it calls AIP bootcamps — condensed, hands-on sessions where a prospective customer's own engineers build a working prototype on their own data, live, typically within days. The pitch is proven or disproven fast, with a tangible result the customer can see immediately, instead of a slide deck promising results after a lengthy implementation. That compressed sales cycle is a meaningful part of why Palantir's commercial customer count and average contract size have both been climbing at the same time — a combination that is unusually hard to pull off.

The Numbers: Growth at an Extreme Price

The performance since the AIP launch has been extraordinary, and it has accelerated rather than faded as the AI-stock cycle has matured. Total company revenue growth reached roughly 85% year-over-year in the most recent quarter, and management's own guidance for full-year 2026 revenue — in the range of $7.65 to $7.66 billion — implies growth of approximately 71% year-over-year for the full year. For a company of Palantir's size, an acceleration in growth rate is the exception, not the rule.

PALANTIR REVENUE GROWTH (YoY)

Latest quarter        ████████████████████████████████████  ~85%
FY2026 guidance        █████████████████████████████         ~71%

The stock has moved accordingly. Palantir traded near $6 per share in 2022, when the market treated it as an unprofitable, government-dependent story stock. Today it trades around $132 — roughly a 22x increase — and the company's market capitalisation stands at approximately $309 billion, making it one of the fifty most valuable public companies on earth.

PALANTIR STOCK PRICE

2022    ▓                                              ~$6
Today   ████████████████████████████████████████████  ~$132   (~22x)

That performance comes at a price that only makes sense if the growth continues for a very long time. Palantir currently trades at approximately 131 times trailing earnings and roughly 54 times sales — multiples that price in years of continued, near-flawless execution. A forward P/E closer to 80x reflects expected earnings growth, but even on a forward basis, Palantir is priced as one of the most expensive large-cap stocks in the market by a wide margin. Worth noting: despite the enormous market cap, Palantir shares are actually down roughly a quarter from where 2026 began — a reminder that even the market's favourite growth stories can carry real volatility once a valuation gets this stretched.

The Controversies

A research note on Palantir that skips the controversies isn't research — it's marketing. The criticisms are specific, current, and well documented, not vague "surveillance" hand-waving.

The most concrete and current flashpoint is ImmigrationOS, internally referred to as "ELITE" — a $30 million Palantir contract with US Immigration and Customs Enforcement running through September 2027. According to reporting from Fortune and the American Immigration Council, the platform ingests data from Medicaid and other government databases to generate "leads" and dossiers on individuals ICE believes may be deportable — turning administrative and health-programme data into an enforcement tool, which is precisely the use case digital-rights groups warn about when they discuss "mission creep" in government data systems.

That contract sits inside a much larger pattern: Palantir has secured more than $900 million in US federal contracts in 2026 alone. The company's ICE and broader immigration-enforcement work has drawn sustained criticism from the Electronic Frontier Foundation, the ACLU, and — on the UK side — Medact, which has raised concerns about Palantir's role inside NHS data systems. There is also a live conflict-of-interest question: Stephen Miller, the chief architect of the current US administration's immigration policy, has been reported to hold a financial stake in Palantir; the White House has stated Miller "fully divested" his holdings early in the administration and that no conflict exists. Both the reporting and the White House's response are on the record, and readers should weigh them accordingly — this is an unresolved public debate, not a settled fact in either direction.

None of this is disqualifying for an equity research report — companies operate in the real world, and defence and intelligence work has always involved this kind of tension. But it is directly relevant to the investment case: regulatory and reputational risk here is structural, tied to the core of what the government side of the business does, not an unrelated tail risk sitting off to the side.

Reading the Controversy as an Investor

It's tempting to treat the ethics debate and the investment case as separate questions — vote with your conscience on one, with your portfolio on the other. For Palantir specifically, that separation doesn't fully hold, and it's worth being explicit about the three concrete ways the controversy translates into financial risk, not just reputational discomfort.

First, contract concentration risk. A meaningful share of Palantir's revenue depends on a relatively small number of government relationships that can change with a single administration, a single budget cycle, or a single high-profile lawsuit. Immigration enforcement work is politically contested in a way that, say, air-traffic-control software is not — a change in administration or a successful legal challenge to a program like ImmigrationOS is a plausible, not hypothetical, revenue risk.

Second, index and mandate exclusion risk. A growing share of institutional capital sits inside ESG-screened or values-aligned funds that explicitly exclude companies with material surveillance or human-rights controversy exposure. As those criticisms accumulate in mainstream reporting, the pool of institutional capital willing to hold Palantir at any price can shrink even if the underlying business keeps compounding — a real technical dynamic, not just a moral one.

Third, headline risk around a stretched multiple. At 131 times earnings, Palantir's stock has very little room for a negative surprise of any kind — and a major new investigative report or a court ruling against a program like ImmigrationOS is exactly the kind of headline that can trigger a sharp repricing in a stock already priced for perfection, independent of whether the underlying financials actually changed.

None of this tells you whether to own the stock. It tells you that "the ethics are a separate issue from the investment case" is not quite true for this particular company — the two are more tightly coupled here than for almost any other large-cap name in the market.

The Bull Case vs. The Bear Case

The bull case: Palantir may be the only company with genuinely proven AI deployment inside the most sensitive, highest-security government and enterprise environments on earth. That track record is extraordinarily hard to replicate — a new entrant cannot simply build classified-environment credibility from a standing start, no matter how good its models are. Once Palantir's software is embedded inside an agency's or an enterprise's core operations, switching costs are enormous; institutions do not casually rip out systems that manage battlefield logistics or hospital data. On top of that entrenched government base, the commercial AIP business is compounding fast and expanding the customer pool well beyond the original defence-and-intelligence core.

The bear case: At roughly 131 times trailing earnings and 54 times sales, Palantir is priced for a level of sustained, flawless execution that very few companies in history have delivered for long. A meaningful share of revenue still depends on government contracts and political relationships that can shift with a single administration change or budget cycle. And the ethical and regulatory risks are not a hypothetical, distant scenario — they are live, current, and directly tied to named contracts (ImmigrationOS), specific dollar figures ($30 million, $900 million+), and ongoing advocacy-group and press scrutiny that could plausibly escalate into real regulatory or contractual consequences.

The Bottom Line

Palantir is not a company that fits neatly into a single narrative, and it shouldn't be forced into one. The growth is real and, unusually for a company its size, still accelerating. The technology genuinely solves a hard problem — fragmented data inside the world's most complex and sensitive institutions — that most competitors either can't or won't touch. The valuation prices in years of continued dominance, which is a real risk on its own regardless of what you think about the business. And the controversies are not noise to be filtered out of the investment case; they are structurally connected to the same government relationships that make the business defensible in the first place.

Palantir earns the label "most controversial tech company in the world" honestly — not because the business is fake, but because the same qualities that make it a formidable, hard-to-replace vendor to governments are exactly what make its critics uneasy. Neither the bulls nor the critics are wrong. They are describing the same company from different vantage points, and the honest version of this stock's story requires holding both.

---

This article is for educational purposes only and does not constitute financial advice.

TFV-DATA:{"category":"Technology · Data & AI","reportNo":"2026-022","titleAccent":"Most Controversial","subtitle":"Funded by the CIA. Now a $309 billion AI bet. The business is real, the growth is real, and the ethics questions are not a footnote.","date":"July 10, 2026","exhibitLabel":"Palantir (PLTR) · Market Cap","exhibitValue":"~$309B","exhibitChange":"~22x since 2022"}