Europe Buys Its Own Defense AI: Helsing Raises $1.8 Billion at an $18 Billion Valuation

A round that grew while nobody was looking

On July 13, the Munich-based defense company Helsing announced it had closed a Series E of $1.8 billion at an $18 billion valuation — the largest funding round a European defense startup has ever raised.

The most telling detail is not the headline number but how it got there. The round was first reported in May at $1.2 billion. By the time it closed, the check had grown by half a billion dollars at the same valuation. That is not a company negotiating harder; that is a company turning people away and then deciding not to. Helsing described the round as heavily oversubscribed, and Defense News reported that demand "significantly" exceeded the allocation available.

Dragoneer led, with Lightspeed as co-lead. The new names on the cap table are the story within the story: JPMorgan Chase, the growth arm of Goldman Sachs Alternatives, and the Canada Pension Plan Investment Board, alongside General Catalyst, Iconiq, and others. Existing backers Prima Materia, Accel, and Greenoaks followed on. The valuation is up from roughly €12 billion a little over a year ago.

When a Canadian public pension fund and two American investment banks are buying into a European weapons-software company at eighteen billion dollars, the asset class has finished arguing with itself.

What Helsing actually builds

It is worth being precise here, because "AI defense company" is a phrase that has been used to describe everything from genuine autonomy stacks to dashboards with a chatbot bolted on.

Helsing was founded in 2021 by Gundbert Scherf, Torsten Reil, and Niklas Köhler, and now employs around 900 people across Germany, the UK, France, and the Baltics. Its portfolio, as described by SiliconANGLE and Defense News, spans software and hardware:

• Altra, a battlefield software layer that aggregates drone and sensor feeds into a real-time picture of an area.
• HX-2, a strike drone that has already been supplied to Ukraine.
• CA-1 Europa, an autonomous jet aircraft — roughly 36 feet long and four tons — explicitly designed for high-volume manufacturing rather than artisanal defense-primes production.
• Centaur, the reinforcement-learning autonomy platform underneath, which SiliconANGLE reports was tested flying a fighter jet over the Baltic Sea last year.
• SG-1 Fathom, a small autonomous submarine that runs neural networks onboard to collect underwater data.

The company has also been buying and building industrial capacity, not just writing code: it acquired the aircraft manufacturer Grob, runs factories in the UK and Germany, and has a third under construction. It works alongside established primes including Rheinmetall, Kongsberg, and Saab.

That last point matters more than it sounds. The failure mode for defense software startups is producing an impressive demo that no procurement office can integrate. Helsing's answer has been to own the metal — to sell aircraft and drones and submarines with the AI inside them, rather than selling AI to someone who makes aircraft.

The comparison everyone is making, and its problem

Helsing is now routinely framed as Europe's answer to Anduril. Both companies reject the idea that autonomy should be a feature retrofitted onto legacy platforms; both raise venture money at software multiples to build hardware.

But the framing flatters Helsing. Anduril is valued at roughly $61 billion — more than three times Helsing's new mark. Europe's largest defense startup is, in the venture arithmetic that both companies have chosen to be judged by, still a challenger with a fraction of the balance sheet.

What Helsing has that Anduril structurally cannot is Europeanness. Co-CEO Torsten Reil noted the company remained about 80 percent European-owned before this round expanded — a number he volunteered, which tells you it is a selling point rather than an accident. The board is a sovereignty statement in itself: co-chaired by Spotify founder Daniel Ek and former Airbus CEO Tom Enders, with retired NATO commander Gen. Denis Mercier among its members.

Helsing is not merely a company that happens to be in Munich. It is a bet that European governments, having spent 2025 and 2026 discovering exactly how contingent American security guarantees can feel, will pay a premium for a capability they do not have to ask permission to use.

Hype versus real

The skeptical read is straightforward, and it deserves airtime.

First, valuation is not revenue. Neither the funding announcement nor the coverage of it discloses Helsing's revenue, order book, or margins, and no such figures have been confirmed. An $18 billion valuation on a five-year-old company is a claim about the future of European defense budgets, not a measurement of the present.

Second, this is a sector where the money arrives faster than the deployments. Helsing's round sits alongside Quantum Systems' $1.2 billion Series D at an $8 billion valuation and Stark's €500 million raise in June. That is an enormous amount of capital chasing a procurement system famously hostile to newcomers. European defense ministries do not buy like enterprise software teams. They buy slowly, in national silos, with parliamentary oversight.

Third — and this is the one nobody in the press release wants to name — the HX-2's deployment to Ukraine is simultaneously Helsing's strongest proof point and its most uncomfortable one. It is real combat validation, which is exactly what separates this company from a slide deck. It is also a reminder that "AI defense" is not an abstraction about compute and reinforcement learning. The Centaur autonomy stack flying a fighter over the Baltic is a research milestone. The same lineage of software in a strike drone is a weapon.

The honest version of the bull case has to hold both.

What this signals about the AI market

Step back from defense and the round says something about where AI capital is going in mid-2026.

The dominant story of the past year has been infrastructure: gigawatts, memory shortages, custom silicon, data center leases measured in decades. Helsing is a different shape of bet entirely. Nobody is funding it because it will train a frontier model. They are funding it because it applies known techniques — reinforcement learning, sensor fusion, onboard inference — to a domain with a captive, state-funded, politically urgent buyer.

That is the applied-AI thesis in its purest form, and the fact that it now clears $18 billion in Europe suggests investors have stopped believing that all the value accrues to whoever has the biggest model. Some of it accrues to whoever owns the drone.

Whether that thesis survives contact with actual European procurement is the open question. The capital has already voted.

The takeaway

Helsing's $1.8 billion is the largest defense-tech round Europe has ever produced, and the oversubscription — from $1.2 billion in May to $1.8 billion at close — tells you the demand was not manufactured. The company has real hardware, real factories, and real combat deployment in Ukraine, which puts it well past the demo stage that most "AI for defense" pitches never escape.

What it does not yet have, at least publicly, is disclosed revenue to justify an $18 billion mark, or evidence that Europe's fragmented procurement can absorb capability at the speed venture capital expects returns. Those are not small gaps. Treat the valuation as a forecast of European defense spending and strategic anxiety, not as a scoreboard of what Helsing has already sold.

The number is confirmed. The thesis is a bet.