Palantir’s wild valuation flags hyper-prime status

Forgive U.S. military contractors their god complex: after all, their resurrection and rebirth followed what became known as the “last supper”. On July 21, 1993, Deputy Secretary of Defense William Perry told a meeting of industry CEOs that, with the Cold War over and spending cuts incoming, they would have to consolidate to survive. The resulting decades of M&A created a select group dubbed the “primes”, the biggest of which – RTX RTX, Lockheed Martin LMT and Northrop Grumman NOC – are collectively worth $400 billion. Their grip on the military supply chain has subsequently caused plenty of angst. Yet in 2025, a single defense-focused software upstart, Palantir Technologies PLTR, has grown at lightning speed to a valuation almost as big as all three.

Formed in 2003 by PayPal co-founder Peter Thiel, Palantir did not achieve this feat through sheer scale. The company, which sells subscriptions to defence and organisation-management software incorporating artificial intelligence, is expected to generate $4.2 billion of revenue in 2025, according to analysts polled by Visible Alpha. That’s way off the $67 billion average forecast for RTX, Lockheed and Northrop. But while they trade on average at only around 2 times forecast 2025 sales, Palantir’s $372 billion market capitalisation weighs in at nearly 90 times. Even among a set of 2,000 sizable U.S. companies that have ever traded above a 70 times multiple since the turn of the century, that makes it one of the most expensive in recent history, reckons Trivariate Research.

Part of this may well be its meme-like status. Retail investors poured $1.2 billion into Palantir stock in July. Even among this loftily valued crew, though, the company stands out. Tesla TSLA – to pick one highly valued stock that largely validated its hype – notched a peak multiple of around 20 times sales in 2010. For Palantir to get even to that level, its revenue would have to grow 35% a year until 2030.

It’s on track to get there right now, at least. Palantir’s sales rose by 48% year-over-year in its most recent quarter. It may not manufacture guns, bombs or tanks. But academic philosopher-turned-CEO Alex Karp can already boast of being the supplier of choice for the new front of superpower warfare: software. Palantir is the prime player on nine U.S. Army “programs of record”, or committed funding streams. This summer alone, the company won a $10 billion, decade-long deal to consolidate 75 different services agreements with the Army. Palantir’s recently expanded $1.3 billion contract for “Maven Smart Systems” is a case in point of how its influence could grow.

At the risk of rehashing Sun Tzu, military operations have always been about command and control: having sufficient oversight of military resources, supply chains and enemy whereabouts to quickly and efficiently make decisions. The rise of artificial intelligence-enabled drones and cyber warfare confronts generals with rapidly increasing complexity. Maven hoovers up files, emails and imaging and blends them with satellite, troop movement and supply data to provide a clearer lay of the land.

Palantir had already made inroads with senior staff with its Gotham software, first deployed over a decade ago to help U.S. personnel in Afghanistan identify roadside bombs. Many in attendance at July’s LANDEURO defence conference in Germany raved about it. Rivals, meanwhile, talked of Palantir’s obvious objective: to provide not just various products, but to become the de facto “mainframe” that can connect and centralise operations. RTX and Lockheed’s control of intellectual property once gave them a stranglehold on certain types of irreplaceable high-tech weaponry, boosting their valuations. Palantir’s software may end up hardwired into the Pentagon to an even greater degree.

The same goes for Western allies, too. In March, NATO also signed up to Maven. For at least the next five years, Palantir’s product will be the platform for bringing together all combined information from the 32-nation alliance.

The scuttlebutt at LANDEURO, though, was that non-U.S. staff are not all quite so enamored of Maven as their American peers. The problem is simple. For Palantir’s extremely expensive product to work to its full potential, every NATO nation needs to share its data with the company and Washington. Amid a belligerent U.S. turn on trade and even territorial claims on Canada and Greenland, not all of them may want to. Reciprocity is wanting, too. Washington’s own Government Accountability Office said in April that overly restrictive policies on data sharing by the American military work against the dream of a NATO-wide system.

These constraints may limit future growth for Palantir’s non-U.S. revenue. In the second quarter of 2024, 36% of group sales came from outside the United States. In the same period this year, that proportion shrank to 27%.

For Karp and Thiel, this may not matter. Uncle Sam is still the world’s biggest defence spender. Moreover, its AI-enabled data product for the private sector, dubbed Foundry, has won glowing testimonials from companies like Hertz, United Airlines and Walgreens. Karp has bombastically predicted that non-government sales, which last year were 45% of overall revenue, will grow tenfold by 2030 – an annual jump of 60%. He managed a 93% year-on-year rate of U.S. commercial sales expansion in the most recent quarter.

Tot it up, and Morningstar analysts suggest that Palantir may take a chunky 3% of a global software market that could be worth $1.6 trillion by 2033. The company’s current valuation sits at a 7.6 times multiple of the implied $48 billion of revenue, compared with 5.4 for Microsoft MSFT relative to the same year.

Meanwhile, choosing to forego Palantir’s capabilities may become increasingly costly and inconvenient for uneasy military allies. The Pentagon still leads the Western alliance; not being on its operating system may not be tenable forever. As a recent 15% mini-slump in the stock implies, investors have priced shares to perfection. But Karp has a not-entirely-implausible path to dominance that the original primes can only envy.

Follow George Hay on Bluesky and LinkedIn.