Signals Inbox·July 21, 2026·Defense Tech

Palantir vs Anduril: who is ahead?

Palantir is ahead today on revenue, profit, military adoption and financial power. Anduril is the more credible challenger, with the stronger position in autonomous systems and the bigger bet on how future wars will be fought.

Q1Why do people keep comparing Palantir and Anduril?

People compare Palantir and Anduril because both want to become the digital backbone of the Western military. Palantir starts with data and decisions. Anduril starts with autonomous machines. Their products increasingly meet in the middle.

Palantir’s software connects intelligence, logistics, targeting, personnel and operational workflows. Anduril’s Lattice platform connects sensors, drones, interceptors, aircraft and other autonomous systems. Both companies want military customers to build future systems around their architecture rather than treat them as another replaceable supplier.

That overlap has become much clearer lately. The U.S. Army created broad enterprise purchasing vehicles for both companies, then selected an architecture combining Anduril Lattice and Palantir Foundry for Next Generation Command and Control, or NGC2. They now compete for the same control point while also depending on each other inside major programs.

So the rivalry is bigger than software versus hardware. Palantir wants to organize the military’s information and decisions. Anduril wants to organize those decisions and supply many of the machines carrying them out.

Q2Why is choosing between Palantir and Anduril so difficult?

Palantir is winning the business that exists today, while Anduril is winning more of the programs that could reshape defense tomorrow. That is the tension behind almost every comparison.

Palantir gives us audited revenue, margins, customer expansion and cash flow. Anduril gives us fighter-aircraft awards, autonomous submarines, missile programs and enormous factory plans. One company can show what its advantage already earns. The other can show what its advantage may eventually control.

Their contract values also mean different things. A Palantir software deployment can carry very high margins without requiring a new production line. An Anduril program may generate much more revenue per customer while also consuming capital, components, labor and years of testing.

That leaves two questions. Who has built the stronger business now? Palantir. Who is making the more credible attempt to become a new weapons prime? Anduril.

Q3Who is bigger today, Palantir or Anduril?

Palantir is currently operating at roughly twice Anduril’s revenue scale and far greater financial scale. Palantir reported $4.48 billion of revenue for 2025. Anduril disclosed $2.2 billion during its latest funding round, giving Palantir a 2.03-to-one advantage.

The more important gap appears after revenue. Palantir produced an 82% gross margin, $1.41 billion of operating income and $1.63 billion of net income in 2025. Anduril has released no comparable margin or profit figure.

Palantir then generated $899 million of operating cash in the first quarter of 2026 alone. That single quarter of cash generation equalled about 41% of Anduril’s entire previous-year revenue.

Anduril carries the heavier engineering challenge, so a direct margin comparison would be unfair. It designs aircraft, missiles, underwater vehicles, interceptors and factories. Still, the current scoreboard is lopsided. Palantir already has the economics of a highly scalable software platform. Anduril still has to prove the economics of a vertically integrated defense manufacturer.

Palantir and Anduril financial comparison

Q4Is Anduril growing faster than Palantir?

Anduril grew faster in 2025, but Palantir is currently adding more revenue and has accelerated sharply. Anduril more than doubled to $2.2 billion, while Palantir grew 56% to $4.48 billion.

The percentage gap looks decisive until we examine the dollars. Palantir added $1.61 billion of revenue during the year, equivalent to roughly 73% of Anduril’s entire business. Anduril may have grown faster, but Palantir almost added another Anduril-sized company in two years.

Palantir has since accelerated further. Its first-quarter revenue rose 85% to $1.63 billion, including 84% growth in U.S. government revenue and 133% growth in U.S. commercial revenue. Management now expects about $7.66 billion for the full year, around 71% above 2025.

Even another doubling would leave Anduril behind that forecast. Revenue would reach $4.4 billion, still around $3.3 billion below Palantir. Closing the gap requires several more years of exceptional execution, particularly because manufacturing usually becomes harder as volumes rise.

Anduril wins the historical growth comparison. Palantir currently has the stronger combination of growth rate, absolute revenue added and margin.

Q5Is Palantir’s revenue lead mostly coming from outside defense?

Palantir’s government business alone was already slightly larger than all of Anduril. Palantir generated $2.40 billion from government customers in 2025, about 9% more than Anduril’s total revenue.

That adjustment makes the defense race much closer than the headline two-to-one company comparison suggests. Palantir earned another $2.07 billion from commercial customers, including manufacturers, healthcare groups, financial institutions and energy companies. Removing that commercial business cuts most of its apparent size advantage.

One caveat: Palantir’s government category includes civilian agencies and foreign governments. Anduril does not disclose how its revenue divides between development work, production, software, services, U.S. customers and allied governments.

Still, Anduril has reached approximately the scale of Palantir’s complete government segment despite being founded 14 years later. That is one of the strongest facts in Anduril’s favor. Palantir leads the overall company race comfortably, but the directly contested defense business is already close.

Q6Does Anduril’s $20 billion Army contract put it ahead of Palantir?

Anduril’s $20 billion Army vehicle gives it the broader procurement channel, but the Army has not promised Anduril $20 billion. The figure sets a maximum purchasing ceiling over ten years. Actual orders could be far lower.

The same warning applies to Palantir’s $10 billion Army agreement. The Army consolidated 75 contracts into one framework allowing Palantir products to be purchased with pre-negotiated pricing and volume discounts. Its announcement explicitly said the ceiling represented no specific obligation.

Anduril’s agreement consolidates more than 120 procurement actions and covers software, hardware, compute infrastructure, data systems and support. That broader scope naturally produces a larger ceiling than Palantir’s mainly software-focused vehicle.

Anduril still wins this comparison. More products can now be purchased through one route, negotiations should move faster, and units have a simpler path from experimentation to deployment. That is a major distribution advantage.

The mistake would be treating the $20 billion as backlog. It is a wider sales channel. It is not future revenue.

U.S. Army enterprise agreements compared

Q7Who currently controls the military operating system?

Palantir currently owns more of the military data layer, while Anduril has become the stronger control layer for autonomous systems. Palantir’s position is broader today.

Maven, Foundry, Gotham, Army Vantage and the Army Intelligence Data Platform connect information from existing military systems and turn it into operational workflows. Palantir can sit above sensors, databases and weapons made by many different suppliers.

The latest evidence goes well beyond pilot programs. A Pentagon memo reviewed by Reuters moved Maven toward a permanent program-of-record structure. NATO has also expanded Maven across Allied Command Operations. During the Steadfast Deterrence exercise, NATO’s Joint Warfare Centre said all Allied Command Operations headquarters were using the system and nearly 550 staff used it at the centre alone.

Recent reporting from the Financial Times placed Maven inside live U.S. targeting and planning operations at an unusually high strike tempo. Whatever one thinks about those operations, this is stronger proof of military dependence than another demonstration or contract announcement.

Anduril’s Lattice has a different strength. It can detect objects, fuse sensor feeds, assign missions and coordinate autonomous platforms at the edge. That makes it particularly valuable when the military wants machines to react together rather than merely give commanders a common picture.

Today, Palantir reaches more of the organization. Anduril reaches deeper into autonomous action.

Q8Did Anduril’s NGC2 win change the race?

Yes, NGC2 moved the software race toward Anduril. The Army chose Anduril to lead the common data baseline for its new command-and-control architecture, giving Lattice its most important software role so far.

This was earned through operational experimentation rather than awarded from a presentation. The Army tested competing approaches across the 4th and 25th Infantry Divisions, including exercises under simulated cyber and electromagnetic pressure.

Anduril said its team expanded the data mesh beyond 65 tactical edge nodes and completed a cross-service sensor-to-effector chain without manually re-entering data. The Army’s subsequent decision gives that company-reported result much more weight.

Palantir remains inside the architecture. The selected baseline combines Lattice and Foundry, with Anduril leading and Palantir supplying part of the edge-to-cloud data mesh. The Army also intends to preserve competition rather than hand the complete system to one supplier.

Even so, lead integrator is the better seat. Anduril can influence interfaces, implementation choices and the surrounding vendor ecosystem. Before NGC2, Lattice looked strongest when attached to Anduril hardware. It now has a credible claim to leading a multi-vendor Army software architecture.

Palantir still has the larger military software footprint. Anduril has finally won a control point that can challenge it.

Q9Whose technology has worked better in real operations?

Palantir has the cleaner operational record today. Its software is already used across military, intelligence and NATO workflows, while Anduril’s hardest products are still working through reliability problems.

Palantir’s advantage comes from repetition. Maven is being institutionalized across the Pentagon. NATO is using its own deployment across command headquarters and exercises. The Army selected Palantir after competition for the $178 million TITAN prototype program, which connects space, aerial and terrestrial sensor information for targeting.

Anduril also has meaningful deployments. Customs and Border Protection has used its autonomous surveillance towers for more than seven years. Australia said all three Ghost Shark prototypes arrived on budget and ahead of schedule. The U.S. military has purchased Roadrunner-M interceptors and Pulsar electronic-warfare systems through a contract worth roughly $250 million.

The weak point is consistency across newer systems. The Wall Street Journal reported that more than a dozen unmanned boats using Lattice stopped during a Navy exercise and required overnight recovery. It also reported setbacks involving the Fury aircraft, an Anvil counter-drone test and drones used in Ukraine.

Anduril disputed parts of the account. It said the boat failure came from software on the vessels, that the bug was fixed, and that the systems later completed autonomous missions. Testing failures are also normal in weapons development.

The pattern still counts. Palantir’s failures usually concern integration, implementation or software performance. Anduril can face those same problems alongside engines, airframes, fires, supply chains and vehicle recovery. The physical consequences are harder to contain.

Q10Can Anduril actually manufacture weapons at scale?

Anduril has become a real manufacturer, but it has not yet proved high-volume production across several weapon families. Its recent progress is much stronger than another factory promise.

The FQ-44 autonomous fighter provides the clearest sequence. The Air Force selected Anduril for prototype work in April 2024. The aircraft began flight testing in late 2025. In June 2026, Anduril received an initial production contract with a route toward additional aircraft lots.

Moving from prototype selection to production in roughly 26 months is unusually fast for a fighter-class aircraft. The contract also gives Arsenal-1 a genuine anchor program rather than leaving the Ohio factory dependent on hypothetical future orders.

Ghost Shark offers another useful test. Australia moved from co-development to three prototypes and then a five-year A$1.7 billion acquisition, support and development contract in about three years. A dedicated Sydney production facility opened shortly afterwards.

Anduril is now adding Barracuda cruise missiles, rocket motors and other systems to that manufacturing plan. It has also agreed with Poland’s PGZ to build surface-launched Barracuda-500M missiles in Europe.

The unresolved part is volume. Arsenal-1 is planned to exceed five million square feet and require close to $1 billion of investment. Producing one aircraft family efficiently would already be difficult. Anduril wants common factories and software to handle aircraft, missiles, drones and other autonomous systems.

The company has crossed from prototypes into early production. The next test is whether output can rise without quality, schedules and margins breaking down.

Q11Which company is solving defense technology’s hardest problem?

Palantir is fixing the military’s immediate problem, while Anduril is attacking the larger long-term one. Palantir has stronger proof so far.

Military organizations already possess thousands of databases, sensors, applications and weapons that struggle to exchange information. Palantir makes those fragmented systems more useful without waiting for every aircraft, radio or database to be replaced.

That bottleneck exists now. Maven, NATO adoption, TITAN, Army Vantage and Palantir’s place inside NGC2 all show customers repeatedly paying to connect old and new systems.

Anduril is targeting the industrial shortage behind a future high-intensity war. Western militaries may need far more interceptors, autonomous aircraft, missiles, drones and underwater vehicles than traditional factories can supply. Arsenal-1, Barracuda, FQ-44 and Ghost Shark all follow that thesis.

Anduril could eventually solve the more important constraint. A military with perfect data still loses if it lacks enough weapons. For now, the factory network and production system remain under construction, while Palantir’s integration layer is already deployed and generating billions in revenue.

Palantir wins on evidence. Anduril wins on the scale of the problem it has chosen.

Q12Who has the better customers and distribution?

Palantir has the stronger customer machine today. It reaches more organizations, expands faster inside existing accounts and can distribute software without building a new supply chain in every country.

Palantir ended the first quarter with 1,007 customers, up 31% over twelve months. It closed 206 deals worth at least $1 million during that quarter, including 47 worth at least $10 million.

Existing customers are also spending far more. Net dollar retention reached 150%, meaning the previous customer cohort generated around 50% more revenue than it had one year earlier. The average annual revenue from Palantir’s twenty largest customers rose 45% to $93.9 million during 2025.

That combination is difficult to beat: more customers, larger contracts and sharp expansion inside existing deployments. Software can also spread across departments and countries without Palantir manufacturing another physical unit.

Anduril has fewer visible relationships, but several are becoming deeper. Customs and Border Protection recently ordered more than 200 additional extended-range towers through a $363 million contract after seven years of working with the company. Australia has moved Ghost Shark into a multi-year fleet program. The Army’s enterprise vehicle gives units a simpler way to buy Anduril products.

Its international distribution is improving too. Local manufacturing agreements with Rheinmetall and Poland’s PGZ make Anduril more credible in European markets where governments want domestic production and sovereign control.

Anduril’s black box is the problem. It does not disclose customer count, retention, contract concentration or repeat-purchase rates. Its large programs look valuable, but we cannot measure whether customer expansion matches Palantir’s.

Q13Which company has the harder advantage to copy?

Palantir is harder to remove today. Anduril may become harder to copy later.

Palantir’s moat grows inside each customer. Data models, permissions, applications, security rules and operating workflows accumulate around its platforms. A competitor must reconstruct that operational logic, not simply move files into another database.

The 150% retention figure supports this. Customers are buying substantially more after deployment, which suggests Palantir becomes more useful as additional teams, data and workflows move onto the platform.

Security accreditation and institutional trust deepen the advantage. Palantir already works across highly classified environments, multinational commands and large government systems. A startup with better software cannot reproduce those approvals and relationships quickly.

Anduril is building a wider moat across autonomy software, sensors, vehicles, test data, factories and supply chains. Reproducing the complete stack would require expertise in artificial intelligence, aerospace, maritime systems, propulsion, weapons integration and high-rate manufacturing.

That breadth also creates risk. Each additional product introduces another specialist competitor and another point where costs or reliability can disappoint. The Air Force has already tested Shield AI’s autonomy software on Anduril’s aircraft, showing that customers want interchangeable components rather than permanent dependence on one vendor.

Palantir’s moat already produces measurable expansion and profit. Anduril’s moat will depend on whether customers keep Lattice, vehicles and manufacturing tied together as programs grow.

Q14Who has more financial power to keep fighting?

Palantir can fund the competition itself; Anduril still needs investors. That gives Palantir much more room for mistakes, acquisitions and long procurement cycles.

Palantir finished the first quarter with $8 billion in cash, Treasury securities and equivalents, with no debt. It generated $899 million of operating cash and $925 million of adjusted free cash flow during the quarter. Full-year adjusted free cash flow guidance sits between $4.2 billion and $4.4 billion.

Anduril’s financing access is extraordinary for a defense manufacturer. It recently raised $5 billion at a $61 billion valuation, nearly doubling its valuation from the previous round. The capital should support factories, inventory, research and production programs for years.

Its strategy simply consumes much more money. Aircraft and missile production requires facilities, tooling, components and workers before the final customer order becomes revenue. Testing setbacks can also delay deliveries while costs continue.

Investor valuations show how differently the market sees the two models. Palantir’s current market capitalization is around $340 billion, approximately 5.6 times Anduril’s latest private valuation. Based on 2025 revenue, that represents roughly 76 times sales for Palantir and 28 times for Anduril.

Palantir deserves a premium for margins, growth and cash generation, but 76 times sales leaves very little room for disappointment. Anduril carries more operating risk. Palantir carries more valuation risk.

The financial winner is clear. Palantir already creates the money required to extend its lead.

Q15Who is winning right now, Palantir or Anduril?

Palantir is winning right now. Its advantage comes from three things Anduril cannot yet match: proven economics, deep operational adoption and the ability to finance growth internally.

Palantir is roughly twice as large, highly profitable and growing faster than most companies ever do at its scale. Its software has spread from isolated deployments into Army platforms, NATO headquarters, targeting workflows and enterprise-wide contracts. Customers are spending 50% more after adoption, and the company is generating close to $1 billion of operating cash per quarter.

Anduril deserves more credit than a standard “promising challenger.” It already has multibillion-dollar revenue, a $20 billion Army purchasing channel, leadership of the NGC2 data baseline, an autonomous fighter entering production and a major Australian undersea program. These wins are arriving repeatedly rather than through one lucky contract.

The difference is what each company still needs to prove. Palantir must keep growing and defend its position inside increasingly open, multi-vendor architectures. Anduril must make aircraft, missiles, underwater vehicles and autonomy software reliable at volume while building factories and acceptable margins at the same time.

Anduril can change the answer. We would need to see several consecutive periods of strong production revenue, improving manufacturing economics, dependable field performance and broader adoption of Lattice beyond Anduril’s own hardware.

Palantir can lock in its lead by keeping Maven and Foundry inside the military’s common architectures, sustaining high customer expansion and converting its current growth into durable government programs.

Palantir still owns the stronger company. Anduril is closer than the revenue gap suggests, but it remains the challenger.

Palantir vs Anduril: who is ahead by criterion?