A new crop of defence-focused venture builders and accelerators are helping early-stage startups break into the tight-knit world of military technology.
As Europe hikes its military spending the race is on to find the next generation of companies that can deliver the required cutting-edge defence tech, including drones, surveillance systems and body armour.
It has spurred a wave of new defence-focused venture builders and accelerator programmes, all looking to give emerging defence companies an alternative route to scale, outside of the traditional system of large military contractors.
One of these is NUNC Capital, a Dutch defence tech venture builder, which last month launched a €20m ($22.7m) initiative to build and scale Ukrainian companies. Bram Oostvogel, NUNC’s CEO, says that the venture building model can generate much more disruptive innovation than the large defence primes.
“The big primes, or OEMs, are typically risk-averse,” say Oostvogel. “They don’t need to act because the whole military industrial complex is built on large or long-term contracts. The scope of work is always written by the MoD [ministry of defence], so there’s not a lot of initiative.”
And for smaller companies, the example of Anduril in the US, which broke into defence contracting by supplying new products that the military found useful, was a “game changer”.
“This is how [innovation] should work. It should come from industry being creative, looking at solutions, which are solving real world problems from the end-user perspective.”
“It’s really up to the industry to be more disruptive and to make products that really make the difference, and not asking the MoD, what do you want me to make?”
At the same time, however, the military establishment itself is setting up some of these startup opportunities. In January the UK MoD’s Defence and Security Accelerator (DASA) partnered with the Janus Consortium, a group of companies involved in dual-use technologies, to launch an accelerator programme for startups. It forms part of the UK’s commitment to Nato’s Diana programme, which scales startup technologies for use across the alliance.
The Janus Consortium is led by IoT Tribe, a company that runs accelerator programmes, partnering with deep tech VC firm Atmos and the six-university SETsquared Partnership. Tanya Suarez, CEO of IoT Tribe, says the accelerator contributes to strengthening Europe’s defence base at a time when this is badly needed. While Europe excels at university research into deep tech, and there is cutting-edge R&D being performed by some of the large primes, it still struggles to bring new technologies to market.
“What we’ve been less good at – but it’s not something that we can’t fix – is the commercialisation [of new technology],” she says.
“So we’ve been perhaps more [effective] on the research side and on the creation of the technology, and less on the exploitation. I think we’ve got a once-in-a-lifetime opportunity to turn the tables on that [now].”
Links to Nato
The DASA accelerator drawns on the relationship with Nato to ensure the solutions put forward fit the alliance’s strategic need.
“Nato is precisely that pan-European forum that provides the opportunity to see what is best in class, and how we can get that out there. So we do talk to generals from Nato to get first-hand information,” says Suarez.
At NUNC Capital, Oostvogel, who served in the Dutch special forces before starting his business career, draws on the links he has established with with important stakeholders, including Nato and large defence contractors. The closed-off defence sector can be hard for startups to break into without insider guidance on how the system works.
“We give [the portfolio companies] a platform, our knowledge, our resources and our experience,” he says.
“We can make product-market fits better than others or quicker than others, because we can test [the technology], for example, in Ukraine. We also have our own distribution channels in our [portfolio], so we can also check [the product] with end users, and we will also check it with primes. So our unknown unknowns are much less than other companies, because we have our own ecosystem.”
Oostvogel founded NUNC Capital in 2013 and has built a portfolio of defence tech companies, including VEARCO, which stops vehicles using non-lethal methods, and PACPROTECT, which makes ballistic protection products.
In Ukraine, the latest venture building initiative will look to invest in aerial, sea and ground drones, as well as electronic warfare systems. The goal is to build companies that are successful in Ukraine but which can then be scaled to Europe and the rest of the world.
Oostvogel says that NUNC Capital does not view the primes as competition. For one thing, they are unlikely to have the same broad attitude towards innovation as a venture builder can have, because they are often public companies, and so need to be more cautious with experimental technology. But the umbrella of a large corporate can also be unsuitable to startup founders.
“The fit for the startups or the scale-ups is quite challenging, because they come in a very stringent, or very corporate, structure, which is killing for most founders, because they want to act quick.”
How defence startups are using the programmes
There are five companies in the January cohort of the DASA accelerator programme. These are: AI Verse, a French startup making synthetic image datasets to train visual AI models; Metahelios, a UK nanotech company making sensors; RVmagnetics, from Slovakia, which makes miniature sensors; and two Finnish startups, EIFys and Winse Power, which make light detection and light-based energy technology, respectively.
As with NUNC Capital’s targets for the venture building programme, the startups are at a technology readiness level (TRL) of four to six, meaning they have either been successfully demoed or are on the verge of reaching that stage, but they are still a way off from being made commercially and at scale. Suarez says the criteria are that “there is some evidence that they work, there is some evidence of a product market fit.”
While the potential for an acquisition by a corporate is always kept in mind, Suarez says this is not the only option open to the startups as they scale. “The M&A option is still on the table, but there are many other opportunities that might be more appropriate,” she says.
She gives the example of corporate partnerships, where the startup works with a larger company but is not acquired by them, allowing for a mutually beneficial exchange of resources.
Crossing the divide
Because the technology is emerging independently of the primes, startups can face hurdles breaking into the industry. For the DASA cohort, forging relationships with primes can significantly speed up the pace at which they scale.
Vladimir Marhefka, vice chairman of RVmagnetics, a Slovakian company that is part of the cohort, says that being affiliated with Nato’s Diana programme has opened doors.
“It significantly helps in the commercial negotiations and in general attraction of our company,” he says. “Both Nato Diana and also the European Space Agency became, for us, very evidently the main entry points that would teach us about the defence world [which is a] very, very close-knit world.”
RVmagnetics makes miniature sensors which can be placed in various materials, such as composites, which are widely used in manufacturing. They detect temperature, pressure, vibrations and other physical conditions, providing data that could help spot faults, cut down on waste, improve quality and monitor performance.
This makes them applicable to several industries. Marhefka says that the company intends to launch defence projects in 2026, having built up relationships with the industry. He believes the accelerator programme made this process much quicker.
“Despite the route we took in advancing the technology to get closer to the market [before joining the programme], we [still] needed to understand how to talk with defence primes and defence subcontractors.”
France’s AI Verse, another startup in the accelerator cohort, which previously partnered with Thales, the French defence prime. Benoit Morisset, its CEO and founder, says that partnerships like that can be mutually beneficial.
“I don’t feel there is a huge gap [that would cause difficulty] to connect two teams from a startup to a prime. It works very nicely,” he says.
“Depending on [which prime] we are talking to, we can also have access to their use case, or at least one part of one level of their use case, which is, for us, very valuable, because we learn the problem they need to solve, and so that can help us drive the development of [our product].”
AI Verse has created an image generation engine to make synthetic datasets for training visual AI systems. If a user wants to train their drone system to target a certain type of tank, for instance, they can input the range of altitudes the drone will fly at and the specifications of its camera. AI Verse’s engine will then create thousands of labelled computer-generated images of this tank relevant to the drone’s vision parameters. It will adjust the terrain, the weather, the lighting and any other variable that is not specified to be fixed, to produce a comprehensive set of pictures to train on.
Morisset founded the company after he worked for Apple on a visual AI project, where he says he realised that manually gathering and labelling images for AI training was a laborious process. While it might be feasible for a company with Apple’s resources, it would be prohibitively time-consuming and costly for a smaller company.
“This whole pipeline—collecting images that fit your use case with the right level of variety, then labelling them while filtering out errors—is, in most cases, a bottleneck for everyone and, in many cases, signals the end of the project itself,” he says.
Morisset says an AI image generation engine is a useful tool for any industry looking to incorporate visual AI, but that it had to pick one specific industry to make inroads in before it could scale. Defence had a number of advantages. For one thing, gathering data manually is especially hard for defence manufacturers, who might encounter legal hurdles to taking lots of drone pictures, for example. The sharing of datasets within and between organisations is also more difficult because of the higher security concerns.
But another attraction was that the defence sector is now expected to grow rapidly in an increasingly uncertain world.
“Sovereignty is a big concept now, and a very clear concept for everybody,” Morisset says, referring to European autonomy from the US’s security umbrella.
“And sovereignty is also applicable to data. And so being able to control your data and share your data without friction and dependency, this is what we bring to the table.”
