Who is the next Palantir?

Why is everyone looking for the next Palantir now?

People are looking for the next Palantir now because defense startups are winning operational programs, not just experimental contracts, while Palantir itself keeps getting bigger.

Anduril generated approximately $2.2 billion in revenue in 2025 and raised $5 billion at a $61 billion valuation in May 2026. In October 2026, an employee share sale reportedly priced the company at about $90 billion. That sale moves shares between holders and brings Anduril no new money. Helsing raised $1.8 billion at an $18 billion valuation in July 2026. Shield AI secured $2 billion in new financing at a $12.7 billion valuation.

The contracts have also changed in kind. In October 2026, the Army gave Anduril a contract worth up to $1.8 billion to scale its command-and-control data layer, and Air announced a separate $200 million award for the same program a week later. Both figures are ceilings or announced values, not money already paid.

Those valuations would have looked absurd when Palantir was founded. For years, venture investors avoided defense because government sales took too long, procurement favored established contractors and startups struggled to turn prototypes into production contracts.

Palantir, meanwhile, has become a harder benchmark: its quarterly revenue almost doubled in a year. Investors are asking which company could gain the same kind of institutional power. That question is much larger than which startup has the best defense technology.

What would being “the next Palantir” actually require?

The next Palantir would need to become a system that governments depend on, rather than a supplier they occasionally buy from.

A company can sell billions of dollars of drones without becoming Palantir-like. Another company may build software that looks almost identical to Palantir’s products but remain too small to shape how major institutions work.

We are looking for a company that keeps gaining responsibility after its first deployment. It should begin with one difficult mission, connect more data and systems over time, and become useful to a growing number of people inside the customer. Each new project should strengthen a common product instead of becoming a separate consulting assignment.

Government access is only part of the test. Palantir proved that technology developed for intelligence agencies and militaries could also run factories, hospitals, airlines, energy networks and financial operations. In the second quarter of 2026, its commercial customers brought in $945 million, almost as much as the $990 million from governments. A successor needs some credible path beyond one military buyer or one weapons program.

What we need to see	Why it matters
Repeated production use	Pilots can disappear without changing how an institution operates
A shared software platform	Separate products do not create the same lock-in or learning effect
Expansion inside customers	Palantir became powerful by taking on more work after entering an organization
Connections to third-party systems	A platform becomes more valuable when competitors also have to plug into it
A market beyond one program	One large contract can create revenue without creating a lasting institution

Why is Palantir so difficult to copy today?

Palantir is harder to copy than its “government software” label suggests, because its customers keep spending more once they are inside the platform.

Palantir reported $1.935 billion of revenue in the second quarter of 2026, up 93% from a year earlier. U.S. commercial revenue reached $764 million, up 149%. U.S. government revenue reached $809 million, up 90%. The company said net dollar retention was 157%, meaning existing customers spent far more than they did a year before.

That is the engine competitors need to reproduce. Palantir enters through one problem and then spreads into more teams, datasets and decisions. In the same quarter it closed 220 deals worth at least $1 million, including 73 worth at least $10 million, for $3.373 billion of total contract value.

The business is also much broader than government intelligence. Worldwide commercial revenue was $945 million in the quarter, up about 110%, against $990 million from government customers. Palantir said its U.S. commercial remaining deal value had reached $6.2 billion.

Palantir achieved that expansion with software economics. Its 10-Q implies a gross margin of about 85% for the quarter; the company’s adjusted figure, which excludes stock-based compensation, is about 86%, so the two numbers measure different things rather than disagree. GAAP operating margin was 47%. In August 2026, Palantir raised its full-year revenue guidance to $8.150 billion to $8.158 billion. That is a forecast, not achieved revenue.

The Ontology, Foundry, Gotham and Apollo are important, but the harder advantage to copy is the loop around them. Palantir learns how organizations operate, turns that knowledge into reusable software and uses the improved product to win larger deployments.

Winning several defense contracts does not recreate that loop.

Is Anduril currently the closest company to Palantir?

Anduril is the closest overall successor because it combines meaningful scale, major government programs and software that connects products built by other companies.

Anduril’s approximately $2.2 billion in 2025 revenue was close to half of Palantir’s $4.5 billion. The comparison is imperfect because Anduril sells expensive physical systems, but it shows that the company has moved far beyond startup-sized government experiments. Anduril does not publish quarterly results, so no like-for-like 2026 comparison is possible.

Its growth has spread across several layers of defense. Anduril began with surveillance and counter-drone systems. It now works on autonomous combat aircraft, underwater vehicles, tactical communications, rocket motors, cruise missiles, submarine components and command-and-control software.

The products reinforce one another. Anduril can sell an aircraft, connect it through Lattice, add data from third-party sensors and later introduce other autonomous systems into the same network. That gives the company several ways to enter a military customer and several more ways to expand.

Recent contracts show that buyers trust Anduril with complete capabilities. The U.S. Air Force moved its collaborative combat aircraft into production. NATO chose Lattice for an allied air-operations data platform. In October 2026, the Army moved Anduril from division-level testing to corps-level fielding of its next command-and-control data layer.

The tension is in how that contract is reported. Headlines said “$1.8 billion,” but only $162.8 million is the initial base. The rest depends on the Army exercising options over five years. It is a strong signal of trust, not $1.8 billion of revenue.

Helsing, Scale AI and Shield AI each beat Anduril on one part of the comparison. None matches its combination of revenue, software, manufacturing and access to major programs.

If you want more recent data on this point, see the latest figures in our Defense Tech report ↓.

Is Anduril’s Lattice now the Army’s command-and-control backbone?

Lattice is close to becoming the Army’s command-and-control backbone, but it is the shared data layer inside a multi-vendor system, not a single operating system that controls everything.

In June 2026, the Army chose Anduril to lead the common data baseline for Next Generation Command and Control. In October 2026, it awarded the follow-on contract: a $162.8 million base to bring that data layer to I Corps, the Pacific-focused corps that commands four divisions, with options that could take the total to $1.8 billion over five years. Defense trade press reported the same base and ceiling as Anduril’s announcement.

The step matters because it moves Lattice from experiments to an operational formation. During an earlier divisional exercise, the network connected 40 applications, 36 data feeds and 37 sensors or effectors across nine hardware formats. Two divisions tested the technology over the past year before the corps award.

Anduril also used Lattice beyond the Army. In August 2026, a Lattice-based battle manager coordinated data across services during the 11-day Valiant Shield exercise in the Pacific. That was a demonstration, not a fielded program.

The architecture is deliberately open. Palantir participated inside the Anduril-led team in the earlier phase, and Air announced its own $200 million award for the same program in October 2026. Anduril controls the integration layer; other companies plug their software into it. That is the most Palantir-like position in the program, and also the reason it is not exclusive.

Lattice has started moving internationally too. NATO selected it for a modular air-command data platform that connects existing allied systems. The comparison still has limits. Palantir models hospitals, factories, supply chains, financial transactions and military operations. Lattice concentrates on sensors, weapons, communications and autonomous systems.

Question	Palantir	Anduril
What does the platform connect?	Data, people, workflows and decisions across whole organizations	Sensors, applications, communications, weapons and autonomous systems
Can it use third-party products?	Yes, this is central to Foundry and Gotham	Yes, the Army NGC2 data layer is built for many vendors
Where does it operate?	Government and many commercial industries	Primarily defense and border security
What strengthens the platform?	More users, data models and operational workflows	More connected units, vehicles, sensors and effectors
What remains unproven?	Continued growth at its present scale	Whether the Army exercises the options and Lattice spreads beyond military operations

Does Anduril’s $3.7 billion submarine bet make it less Palantir-like?

Yes. The submarine bet makes Anduril less Palantir-like financially, but even harder to challenge inside defense.

In October 2026, Anduril said it would invest $3.7 billion of its own capital to build Arsenal-2, a shipyard of more than two million square feet at Sparrows Point, Maryland, to supply parts for U.S. submarines. Separately, the Navy agreed to a contract worth up to $2.9 billion, with payments tied to production results. The announcement added the two together as $6.6 billion. That total mixes company spending with a Navy ceiling; it is not a $6.6 billion contract award.

Aircraft, missiles, submarines and underwater vehicles require factories, materials, testing facilities, inventories and complicated suppliers. Anduril cannot distribute those products at the near-zero marginal cost of software. Its long-term margins should therefore remain well below Palantir’s 85% gross margin.

The manufacturing burden also introduces more ways to fail. A software update can often be corrected quickly. A delayed shipyard, defective rocket motor or unavailable component can disrupt an entire contract and consume large amounts of capital.

Yet hardware gives Anduril an advantage that Palantir never had. Every Anduril vehicle or sensor creates another possible endpoint for Lattice. The company can design the physical product, control the software connecting it and manage the factory producing it.

Recent production steps show why that combination is powerful. Anduril signed agreements designed to support thousands of Barracuda vehicles per year. In September 2026, its Barracuda-M missile passed a third-party open-architecture assessment covering 14 domains in four days; first deliveries are scheduled for early 2027. Earlier, Anduril delivered more than one normal month of Pulsar production in less than a week to two government customers.

Palantir became difficult to remove because its software spread across customer operations. Anduril can create similar dependence through a mixture of software, hardware and manufacturing capacity.

The result will look more like a modern defense prime than an enterprise software company.

Could Helsing become Europe’s Palantir?

Helsing is Europe’s strongest Palantir candidate because it combines defense software with something American companies cannot easily manufacture: European political legitimacy.

European governments increasingly want control over military data, software updates, production and intellectual property. Even a technically strong American platform can face resistance when it becomes too central to national defense.

Helsing was built around that concern. Its early software helped military systems interpret sensor data, identify threats, manage electronic warfare and support battlefield decisions. The company then gained positions inside major European programs.

Saab and Helsing have tested the Centaur artificial intelligence pilot on the Gripen fighter. Helsing is supplying the artificial intelligence for a Eurofighter electronic-warfare upgrade under a contract worth several hundred million euros. It is also involved in the artificial intelligence infrastructure for Europe’s Future Combat Air System.

Helsing has expanded much further into hardware. It received orders covering 6,000 HX-2 strike drones and 4,000 HF-1 drones for Ukraine, and in July 2026 it said thousands of HX-2 systems had been fielded there since the start of the year. That is a company figure; no independent count is public. Its Altra software can coordinate reconnaissance systems, artillery and groups of HX-2 drones. Helsing is also developing the CA-1 Europa autonomous combat aircraft with its Grob Aircraft subsidiary.

The company is reaching outward too. In July 2026, it chose West Virginia for its first U.S. factory. In October 2026, it signed a commercial partnership with Destinus, a European drone and missile maker, and said it would join Destinus’s funding round, without disclosing an amount.

Europe’s fragmented procurement market will make expansion slower than in the United States. Helsing may need to satisfy several governments, industrial partners and national production requirements for each large program.

Its European identity nevertheless creates a protected opening. Helsing does not need to defeat Anduril everywhere. Becoming the preferred software-defined defense company for Europe would already create a very large business.

If you want more recent data on this point, see the latest figures in our Defense Tech report ↓.

Has Helsing’s valuation run ahead of the company?

Helsing’s $18 billion valuation has run ahead of its current revenue, although the business underneath it is clearly real.

Helsing raised $1.8 billion in a Series E in July 2026, at a valuation up from a reported €12 billion in June 2025. The Financial Times reported projected 2026 revenue of approximately €441 million. On that basis, the valuation before the new money was around 32 times projected revenue. Helsing has not published audited 2026 revenue.

That is an aggressive price even for a fast-growing technology company. The same Financial Times comparison, made before Anduril’s October share sale, placed Anduril at roughly 13 times expected revenue and Shield AI at about 21 times.

Helsing’s investors are already pricing in several future successes. The HX-2 needs to turn Ukrainian demand into repeated orders from multiple countries. The Eurofighter work needs to lead to larger roles across European aircraft fleets. Altra must become a common battlefield platform. The CA-1 program must move from development into funded procurement. The U.S. factory must produce at volume.

There is evidence that the drones are gaining real user interest. After reports questioned Ukrainian demand, Helsing said more than six Ukrainian units had requested the HX-2 and that the unit conducting frontline tests had asked for more than 1,000 additional systems. Those figures are company disclosures rather than independent procurement records, but they carry more weight than a laboratory demonstration.

Helsing’s valuation leaves very little room for slow adoption. Several projects must become substantial businesses at the same time.

The company looks expensive today. The technology, contracts and production behind the valuation keep it out of the pure-hype category.

Could Scale AI become the Pentagon’s Palantir?

Scale AI is the best candidate for a Palantir-like business spanning the Pentagon and commercial artificial intelligence, but its government position remains less deeply embedded.

In May 2026, the Pentagon expanded Scale’s enterprise agreement from a $100 million ceiling to a potential $500 million. Defense organizations can use the contract to access Scale’s data infrastructure, model evaluation, engineering services and Donovan platform across classified and unclassified environments.

Scale also leads Thunderforge, which applies artificial intelligence to military planning and operational decisions. A separate Army agreement provides up to $99 million for artificial intelligence research and development.

The pieces resemble Palantir’s model. Scale can organize difficult data, test models, connect them to secure systems and build applications around specific decisions. Donovan gives government teams a controlled environment for deploying specialized artificial intelligence agents.

Scale also has something most defense startups lack: a substantial commercial customer base. Its technology serves artificial intelligence laboratories, large companies, healthcare organizations and other enterprises. That provides a natural route between government and commercial work.

The size of the Pentagon agreements needs careful interpretation. A $500 million ceiling allows agencies to place orders up to that amount; it does not mean Scale has received $500 million in revenue. Scale has not announced a major new defense award since the May expansion, and the company does not publish its defense revenue.

Scale has the right ingredients. It still needs to show that customers keep expanding their use after the initial deployment.

Did Meta make Scale AI a weaker Palantir candidate?

Meta made Scale AI wealthier and less neutral, which weakens Scale’s claim to become the next independent Palantir.

In June 2025, Meta invested $14.3 billion for a 49% economic stake in Scale, valuing the company at more than $29 billion. Scale remained a separate business, but founder Alexandr Wang left the chief executive role to lead Meta’s superintelligence effort, and Jason Droege took over at Scale.

The deal created an obvious concern for Scale’s commercial customers. Artificial intelligence laboratories share sensitive information about their training data, model weaknesses, evaluation results and future development with infrastructure providers. A supplier closely tied to Meta becomes less comfortable for companies competing directly against Meta.

Time reported that major laboratories including OpenAI and Google began moving some work away from Scale after the transaction. Rival data providers subsequently reported stronger demand. No public figure measures how much business Scale lost, so the damage is a clear risk rather than a measured collapse.

The Pentagon may worry less about rivalry between commercial model developers. Meta’s money also gives Scale more resources for secure infrastructure, acquisitions and expensive government deployments. Scale acquired ICG Solutions to deepen its national-security capabilities.

Even so, Palantir’s independence became part of its strategic value. It could work across cloud providers, industries, intelligence agencies and technology partners without belonging to a larger platform company.

Scale can still build a large defense artificial intelligence business. The Meta relationship makes it harder for Scale to occupy the neutral position that helped Palantir spread across competing institutions.

For the Palantir comparison, the Meta transaction was a net negative.

If you want more recent data on this point, see the latest figures in our Defense Tech report ↓.

Could Shield AI own the military autonomy market?

Shield AI could become the main software layer for autonomous military vehicles, although that prize is narrower than the role Palantir plays across entire organizations.

Shield AI’s Hivemind software allows aircraft and other vehicles to operate in contested environments where communications or GPS may be unavailable. The company says Hivemind has piloted 30 classes of vehicles, including F-16s, helicopters, jet-powered drones, boats and ground systems.

The U.S. Air Force awarded Shield AI a production contract to supply Hivemind mission autonomy for the Collaborative Combat Aircraft program. Treating autonomy as a separate production capability is important. It means an aircraft manufacturer can potentially choose Hivemind instead of developing every autonomous behavior internally.

Hivemind is also moving to sea. In July 2026, it ran two Thunder Tiger SeaShark surface vessels as a team in a demonstration in Taiwan. In September 2026, it coordinated two Kraken K3 Scout vessels for a surveillance mission. Both were demonstrations, not production contracts, but they show the software working with other companies’ hulls.

Shield AI is also building its own aircraft. In October 2026, it said commitments to its X-BAT combat jet had reached $400 million: $200 million of U.S. government funding, of which $150 million is new, matched by $200 million of Shield AI’s own capital. First flight tests are planned for later in 2026. The $400 million is not government procurement, and it pulls Shield AI further into hardware.

That creates a credible platform business with a hardware anchor. More vehicle makers may adopt Hivemind, which would produce more operational data, more tested behaviors and greater confidence in the software.

Shield AI remains focused on how machines move and fight. Palantir reaches much further into budgeting, logistics, intelligence, planning and organizational decisions.

Shield AI may become the standard autonomy provider inside military vehicles. That would make it extremely valuable without making it a full Palantir successor.

Is Air, formerly Govini, becoming a real rival to Palantir?

Air is becoming a real, if still small, rival. It looks most like an early Palantir because it turns fragmented defense data into a shared picture for acquisition, production, readiness and now command and control.

The company changed its name from Govini to Air as it expanded beyond defense acquisition. Its Enterprise Readiness platform is designed to connect development, suppliers, production, maintenance, logistics and military operations. Air says its platform is used across programs including the F-35, Patriot, Minuteman, B-52, F-16 and several naval systems, and that more than $140 billion of defense programs are managed through it.

The biggest step came in October 2026, when Air announced a $200 million, five-year contract supporting the Army’s Next Generation Command and Control program, starting with I Corps. That is an announced value, not money obligated so far. It sits beside Anduril’s separate award: Anduril runs the data layer, while Air supplies software that shows commanders supply stocks, unit capacity and force readiness.

There are also narrower examples of real workflow adoption. One Air application reduced supplier due diligence from roughly 120 hours to less than 24 hours. In July 2026, the company secured a $31 million software contract supporting the intercontinental ballistic missile supply chain. Air does not disclose its revenue or margins.

The evolution is strikingly Palantir-like. Air started with difficult government data, built applications around specific decisions and is now working inside the Army’s operational command system.

Vannevar and Primer remain credible alternatives. Vannevar has moved from a Defense Innovation Unit prototype into a production contract with a potential $99 million ceiling. Its software and artificial intelligence agents support intelligence, strategic competition and battlefield decisions. Primer helps defense and intelligence teams process unstructured information, monitor threats and produce traceable analysis inside secure environments.

Air has the broadest institutional ambition of the three. Vannevar appears closer to active military missions, while Primer is strongest in intelligence analysis.

All remain far smaller than Anduril, Helsing or Scale. Their software-first models are worth watching because early Palantir also looked small beside established defense contractors.

If you want more recent data on this point, see the latest figures in our Defense Tech report ↓.

Can any startup match Palantir’s commercial reach and margins?

No current candidate matches Palantir’s combination of commercial reach, customer expansion and software margins.

Commercial customers generated $945 million of Palantir’s $1.935 billion revenue in the second quarter of 2026, almost half. U.S. commercial revenue alone grew 149%. That balance gives Palantir access to budgets far beyond national defense.

A gross margin of about 85% and a 47% GAAP operating margin show that Palantir has escaped the economics of traditional government contracting. Heavy customer involvement has not prevented it from producing mature software margins.

None of the private challengers publishes audited revenue or margins, so every comparison here is a judgment, not a measurement. Scale AI comes closest on commercial reach. It works with major companies and artificial intelligence laboratories, while its government business is growing. Human data work, specialized evaluations and engineering services may keep parts of Scale’s business below Palantir’s margins. The Meta relationship has also damaged its appeal to rival laboratories.

Air, Vannevar and Primer have more software-like economics, but their disclosed scale remains much smaller. Anduril, Helsing and Shield AI can become larger through hardware, although every aircraft, drone, missile or submarine part carries production costs.

The next company with Palantir-like institutional power may look very different financially.

Company	Commercial reach today	Likely economics at scale
Anduril	Very limited outside security and defense	Stronger than traditional defense, well below pure software, with heavy factory spending
Helsing	Almost entirely government and defense	Mixed software and manufacturing margins
Scale AI	Large commercial and growing government presence	Closest candidate to software-platform margins
Shield AI	Mostly defense, with possible industrial autonomy uses	Mixed software and aircraft economics
Air	Concentrated in U.S. defense	Attractive software economics if adoption expands
Vannevar and Primer	Mostly government and national security	Software-like, but current scale remains limited

Which Palantir rival has the strongest moat today?

Anduril has the strongest overall moat because competitors would need to challenge its software, installed systems, factories and government-program relationships at the same time.

Lattice gives Anduril the software layer, now headed for corps-level use in the Army. Aircraft, sensors, missiles and underwater vehicles create a growing installed base. Factories, including the planned Arsenal-2 shipyard, add another barrier. Production contracts give the company operating experience and procurement credibility that newer entrants cannot quickly reproduce.

The layers support one another. A military customer may first buy an autonomous vehicle, later connect outside sensors through Lattice and then add more Anduril products to the same network. Switching suppliers becomes harder as those systems begin working together.

The limit is that this is a defense moat, not a commercial software moat. It depends on defense budgets and on the Army exercising options it has not yet committed to.

Helsing is building a similar position in Europe. Its strongest protection comes from sovereignty, access to European military platforms and the political desire for domestic defense technology.

Shield AI’s moat is narrower but potentially deep. Every additional vehicle class, flight test and autonomous mission makes Hivemind easier to trust and harder for a new software provider to match.

Air has a different advantage. Its readiness graph becomes more useful as it connects additional suppliers, programs, maintenance records and operational requirements. That resembles Palantir’s data and workflow moat more closely than the hardware companies do.

Scale AI has broad technology and customer relationships, but artificial intelligence data, evaluations and agent platforms remain crowded markets. Meta’s involvement has also made neutrality harder to defend.

Company	Strongest moat	Main weakness
Anduril	Software, hardware, manufacturing and program access reinforce one another	Capital intensity and dependence on defense budgets
Helsing	European sovereignty plus growing control of software and hardware	Fragmented procurement and a demanding valuation
Shield AI	Operational autonomy tested across many vehicle classes, now including boats	Narrower role inside the wider defense organization
Air	Proprietary defense data and deeply embedded readiness workflows	Heavy concentration in one government
Scale AI	Commercial reach, data infrastructure and model evaluation	Competitive neutrality after the Meta investment

Who is the next Palantir?

Anduril is the next Palantir.

The comparison works because Anduril is gaining the kind of institutional responsibility that once separated Palantir from ordinary government contractors. The military is asking Anduril to connect systems, lead major programs, produce weapons at scale and provide software through which other companies operate.

As we saw above, the Army is taking Anduril’s data layer from test divisions to I Corps, under a contract with a $162.8 million base and a $1.8 billion ceiling. NATO is beginning to use Lattice as shared infrastructure. Production agreements now cover autonomous aircraft, counter-drone systems, mass-produced strike vehicles and submarine parts.

Anduril will not reproduce Palantir’s financial model. Manufacturing will consume more capital and keep margins lower; the $3.7 billion shipyard makes that plain. Its commercial market also remains small. And Palantir is not standing still: its revenue grew 93% in the second quarter of 2026. Calling Anduril a direct copy would miss the interesting part.

The company is building a different route to the same destination.

Palantir became powerful by controlling the software through which governments understand data and make decisions. Anduril is combining that software role with the machines, weapons and factories through which those decisions are carried out.

Helsing is the strongest answer for Europe. Scale AI remains the best candidate for a government and commercial artificial intelligence platform. Air looks closest to the younger, software-only Palantir, and its Army award makes it one to watch.

Anduril has the clearest path to Palantir-level power. The next Palantir will probably sell more metal, carry lower margins and look much more like a defense manufacturer. Today, it already looks like Anduril.

If you want more recent data on this point, see the latest figures in our Defense Tech report ↓.