Built to connect startups with the defence world, NATO DIANA is already stretching beyond its mandate. Nearly three years in, it shows early results, but it is also revealing why innovation in defence is so hard to accelerate in the first place. This piece examines what’s working, and what isn’t moving fast enough.
The defence landscape is changing quickly. Or at least, it is trying to.
The first signal came in February 2022, when German Chancellor Olaf Scholz announced his ‘Zeitenwende’ – a historical watershed – committing Germany to two shifts: reducing dependencies on potential adversaries and significantly increasing defence spending. Since then, countries across Europe, and the NATO alliance at large, have pledged larger shares of their GDPs to defence budgets, calling simultaneously for faster innovation and shorter procurement cycles.
But progress is slow and start-ups – arguably the ecosystem where some of the most innovative solutions could be found – are still facing significant hurdles to access defence customers.
NATO’s Defense Innovation Accelerator for the North Atlantic (DIANA) became operational in June 2023, aiming to bridge this gap. Nearly three years later, is it holding up to the objectives it set for itself? It is too early to tell? Or are there other signals that show that it is having an impact?
“The tech race is on,” as Jean-Charles Ellermann-Kingombe, NATO Assistant Secretary General for Innovation, Hybrid and Cyber, put it at The Hague Summit in June 2025. Russia, he explained, has compressed its product development cycle to as little as two to twelve weeks; China has deeply integrated its defence industry with its armed forces. Western defence industries, in his words, have been “struggling to keep pace.”
On the supply side, there is no shortage of ambition. Europe alone has 35,000 early-stage tech startups, according to Invest Europe, and investment in the defence-specific subset has grown more than 500% since 2021, according to a McKinsey analysis.
Yet progress remains slow. In Germany, despite the Zeitenwende, procurement cycles remain stubbornly long. The same holds true for other European countries, and while much news coverage defaults to calling this inefficiency, Marcus Roberts, Director of Operations at Janus – a startup accelerator within the DIANA network – offered a more measured read. “Ultimately, there is a human being at the end of that capability,” he told Warsight, “and they need to have a number of assurances that it is going to work the way it is supposed to.” Risk, not bureaucracy, is the primary driver.
The problem is compounded on the startup side. Many companies chase headline investment trends – drone technology being the obvious example – without understanding the operational complexity of the environments those systems must operate in. Investment grew quickly, Roberts argues, and startups followed; but the framework to actually shorten procurement cycles, effectively and efficiently, has not yet been put in place.
The gap, in other words, is not principally one of money or will. It is one of architecture.
Enter NATO DIANA. Launched in 2023, it was built on a recognition that the most consequential emerging technologies are no longer being developed inside defence circles alone, and that the alliance needed a structured way to reach across into the commercial world and pull them in.
Jyoti Hirani-Driver, Acting Managing Director and Chief Operating Officer at NATO DIANA, is clear about what that means in practice. “We cannot rely solely on the same types of technologies, we need to be looking for the next big thing that will provide advantage to our Armed Forces,” she told Warsight, “and that breakthroughs might not exist in defence primes, it might be outside.” Diversifying the defence industrial base, she added, is precisely what DIANA is trying to do by reaching into commercial startup ecosystems.
Dual-use products – goods, software or technology that serve both civilian and military purposes – are central to that bet. As Stacey Crump, Deputy Head of Growth Services at UK Defence Innovation (UKDI), wrote to Warsight, civilian product maturity does not automatically translate into faster military adoption. But company maturity is a different matter: more developed businesses tend to find it easier to innovate, scale, and attract investment. A dual-use strategy can actively help them diversify their revenue streams while addressing adjacent use cases in both markets.
DIANA works through a competitive challenge model. Companies apply against specific defence and security problems – such as sensing and surveillance, energy resilience, secure communications – and those selected enter a funded accelerator programme, gaining access to a network of test centres, mentors, and end users across the alliance. A smaller group is then down selected for a second phase focused on scaling and market access. The diagram below maps the full pathway and programme infrastructure.
One of DIANA’s central aims – and the reason it provides funding across both phases, alongside access to accelerator sites connected to venture capital (VC) networks – is to help companies survive what the defence world calls the ‘valley of death’. This is the long, expensive gap between a promising technology emerging and an actual procurement contract. The logic is deliberate: a company’s commercial business continues to generate revenue while DIANA funding gives it the runway to engage with defence customers, adapt its technology, and build the relationships that procurement requires. The dual-use model, in theory, keeps the startup alive long enough to cross over.
But does it work in practice? Can an accelerator operating across 32 countries, with all the national priorities, procurement cultures, and institutional inertia that implies, genuinely move the needle on the very cycles it was designed to shorten? What does NATO DIANA actually deliver?
Connecting commercial innovators with defence customers is less like matchmaking and more like navigation without a map: both sides are moving, but rarely toward each other.
On the innovator side, many companies arrive with a mature civilian solution but little understanding of how defence operates or what armed forces need in practice. The consequences can be concrete: Ukrainian forces reportedly received warehouses full of drones that had never been tested in operational environments – and so resulted in the vast majority not being used. And even when a startup reaches someone in uniform, it often does not realise, as Roberts noted, that it is talking to the wrong person entirely: “They are not the ones with the buying or the decision-making power.”
On the other side sits a complex mosaic of 32 different procurement systems, most sharing the same structural problem: the end user is not the one writing the cheque. Defence procurement prioritises cost and risk over discovering new suppliers, and after three decades of post-Cold War consolidation, much of that world has become accustomed to a familiar, shrinking industrial ecosystem. Many procurement officials, Roberts argued, still struggle to express their needs in terms that startups – organisations already stretched thin and wary of bureaucratic labyrinths – can readily act on.
As Hirani-Driver put it: “How do we communicate to someone who’s never touched defence before? And how do you navigate a 32-nation defence system, with different procurement cycles and demands?”
The answer begins with the challenges. DIANA identifies genuine capability gaps and their urgency by working with end users across 32 allied countries. Challenge Managers are responsible for translating these demand signals into challenge statements that are precise enough to be actionable, while remaining open to unexpected solutions. Then they translate those needs into publicly issued briefs precise enough to be actionable, but broad enough not to foreclose unexpected solutions. “When defining the challenge, we follow the demand signal,” Veronika Yordanova, DIANA Challenge Manager for Maritime Operations, told Warsight, “but we try to remain relatively broad in our scope because we do not want the language we use to limit the solution we could get.”
What makes it distinct from a standard procurement tender is intent: a tender asks industry to deliver something already specified; a challenge asks the commercial world whether something they are already building might address a need they did not know existed.
The scope of that ambition has grown steadily. DIANA launched with three pilot challenges in 2023 – energy resilience, undersea sensing, and secure information sharing. By 2024 it had expanded to five; by 2026, to ten, spanning advanced communications, autonomy and unmanned systems, biotech and human resilience, critical infrastructure and logistics, and others. The volume of applications has grown in step: from 1,300 in 2023 to 3,600 in 2025, for 150 places.
Down-selecting that field requires more than counting scores. The process involves evaluators from varied backgrounds – technical, commercial, operational – precisely because a solution that appears to belong in one challenge may prove more valuable in another. “For instance, we received a lot of applications for the autonomy challenge,” Yordanova observed, “but as we down-selected we realised that many of them could work in other challenges.”
The challenge is not simply a filter. It is an act of translation, from operational need to commercial opportunity, and back again.
Phase 1 of the programme focuses on accelerating companies. DIANA pairs each startup with the most appropriate accelerator site from a network of 16 across the alliance – matched by challenge type rather than geography – and provides access to, and financial support for, testing across more than 200 test centres.
“The first six months effectively give the innovators the resources and connections they need to navigate defence,” Hirani-Driver explained. Challenge managers and accelerators open doors to potential end users and relevant events where they get a chance to start pitching their idea and refining it. DIANA staff (including serving military personnel) and DIANA accelerator sites provide guidance on running a business, legal considerations, navigating defence and procurement cycles, and access to finance.
“One of the most important tasks we have, as challenge managers, is supporting startups in finding their dual-use business case,” Sinziana Irimia, DIANA Challenge Manager for Critical Infrastructure & Logistics, told Warsight. For companies that have never spoken to a military end user, this phase is critical. But even those with prior experience can find that those conversations led nowhere.
Cogitat, a neurotechnology spin-out of Imperial College London that measures cognitive performance and focus, knew this problem well. It had spoken with the military before but kept hitting the same wall: “We can’t tell you what we need, you just have to build it and we will see.” As Allan Ponniah, Cogitat CEO, put it: “This becomes expensive, time-consuming, and a significant risk for startups who invest money in something that may not be bought eventually.” At DIANA, where similar classified restrictions may apply, the role of the challenge manager is crucial. Yordanova explained that while the classified origins of some capability gaps cannot be disclosed, the broader strategic direction is always present. Innovators set their own goals, with Challenge Managers ensuring those goals are grounded in real use cases and end‑user needs.
myLanguage, a California-based startup that developed fully offline, secure voice-to-voice translation – a tool that works with no internet connection and leaves no data trace – had arrived at defence by a different route, through prior work with the US Coast Guard. But the founders wanted more: “We were looking at the war in Ukraine and we wanted to find a way our technology could serve a real need on the ground there,” Dan Yerelian, COO of myLanguage, told Warsight. DIANA put them in front of customers they could not have reached alone and gave them the resources to refine their product and learn the language of defence procurement.
Astrolight, a Lithuanian tech company developing laser-based optical communications – technology that enables high-speed, jam-resistant, near-undetectable data transfer between satellites, ships, and unmanned systems – had already won a tender with the Lithuanian Navy before applying to DIANA, and understood its technology had defence potential. But understanding a market and knowing how to operate in it are different things. Laurynas Mačiulis, co-founder and CEO, told Warsight “NATO DIANA was a perfect instrument to educate us on how to do business in defence, to get more visibility, better networking, expert advice. Everything combined was the right fit for what we needed at that time.”
DIANA offered these companies both an entry point and a way to explore and open doors to other applications. This is something accelerators are also useful for. Pairing with start-ups is done based on the type of challenge they are responding to and not geographical purpose. “Rather than creating regional hubs [which nations often do], we create a hub of eight innovators from across the NATO alliance that are all focusing on the same things, that all have interoperability, that are now all working together trying to solve similar problems, and therefore can bounce ideas off each other,” Roberts explained.
The results can be unexpected. Cogitat were initially puzzled to find themselves sent to Copenhagen for the human health and performance challenge. “Ironically, for a company that measures focus, we came into it quite unfocused, Ponniah said, “but the trip to Denmark was transformational because in Copenhagen we actually found a company that is likely to become our first defence customer.” DIANA helped them prioritise when, at the point of application, they were still unsure which of their technology’s many potential applications was most relevant to defence.
Where challenge managers and accelerators seem to have had a significant impact for the start-ups interviewed for this article, test centres drew more mixed responses.
For Cogitat, testing at the Estonian Aviation Academy was a genuine inflection point, helping move the company from TRL4 to TRL7 in the defence context, even though their technology was at a much higher readiness level in other sectors, such as medicine. Testing also revealed something important: the technology works better as a co-development layer on top of existing products than as a standalone system.
For myLanguage, by contrast, the test centres were not a good fit: their translation model was already more advanced than the infrastructure on offer. What would have helped them more, Yerelian noted, was direct funding to train their model on additional languages – the kind of expensive, unglamorous R&D that test centre allocations are not designed to cover.
Where DIANA’s first phase casts wide, the second one narrows sharply. “We’re working with a much smaller cohort who needs more attention,” Hirani-Driver explained, “and we are very much hands-on, developing opportunities curated to that innovator, to make sure their technology progresses and that there’s an adoption pathway and possibility of a paid contract at the end.”
Of the three companies interviewed for the purpose of this article, only Astrolight made it through to the second phase of the programme. In practice, this meant DIANA facilitating Astrolight’s participation in NATO exercises REPMUS and DIBAX – two of the alliance’s most significant maritime technology and digital communications demonstrations – giving the company a live operational environment in which to test their laser terminal against real defence conditions and in front of the stakeholders that matter.
myLanguage applied for Phase 2 and did not get through, a decision they found difficult and, in their own words, somewhat opaque in its reasoning. A likely explanation is that Phase 2 favours technologies that require operational testing – ruggedised for field conditions and validated against military infrastructure. Astrolight’s laser communications terminal – which must function between ships at sea, on unmanned systems, and eventually in space – has exactly those requirements.
But getting through the DIANA programme and securing a defence contract are two different things, and the distance between them is where the architecture is most tested.
The first obstacle is structural and familiar: the end user is rarely the decision-maker. A procurement official at a Ministry of Defence may recognise the value of a technology, may even advocate for it internally, but getting it funded requires navigating a competition for budget that has nothing to do with the quality of the solution. “One thing is convincing the customer,” Mačiulis observed, “but then they need to get funding approval within the MOD, and there are thousands of different priorities. How you put your technology on top so it gets attention and gets funding approved is another challenge entirely, and one the Rapid Adoption Service cannot help you solve. It can only kick in once you already have a customer and the funding is confirmed.”
The Rapid Adoption Service (RAS) is DIANA’s contracting mechanism, allowing it to award research, development, and prototyping contracts to innovators on behalf of allied nations and NATO bodies. Once a prototype has been validated, follow-on production or procurement contracts can be awarded through NATO’s support and procurement agencies, without reopening competition. Both Cogitat and myLanguage have moved on to the RAS, with the latter now working with the UK MoD (JHub) on Project Veritas.
The second obstacle is less visible but potentially more consequential: doctrine. Integrating a new technology into military operations requires concepts of operations (CONOPS), training frameworks, and institutional familiarity – processes that move on their own timeline and cannot be compressed by an accelerator. Mačiulis was candid about what this looks like from the inside. “There is still limited knowledge, in the defence market, about how our technology operates. We need to educate the stakeholders,” he told Warsight. “DIANA helped us to do it, but it is a not a process you can realistically do within the programme’s timespan of a year.” A year and a half after completing the programme, Astrolight is still in that process.
Where DIANA may have more immediate impact is investment. “Part of the connections DIANA is making is also introducing startups to like-minded investors,” Hirani-Driver told Warsight. Its Capital Network, which will formally launch on 27 April 2026, aims to connect startups with screened investors aligned with alliance priorities. It reflects a broader logic that accelerators like Janus have been developing in parallel: that surviving the ‘valley of death’ is not simply a question of grant funding, but of building a viable commercial ecosystem around companies that are too early for defence revenue but too defence-adjacent for conventional venture capital.
The DIANA stamp, in other words, is not just a credential. It is a signal, to investors as much as defence customers, that a technology is worth attention.
For Mačiulis, this is not an abstract argument. “You have to survive as a company for many years until you can sell to a defence customer, and you will not get any money from them in the meantime,” he told Warsight. “The funding is essential for innovators to survive this Death Valley.” For Astrolight, the answer to that survival question is the one DIANA’s architects designed for: “The only mainstream of revenue for us today is still commercial. Dual use is very important for us. Otherwise, we wouldn’t have survived.”
Measuring DIANA against the programme’s objectives almost three years in would not be fair, and three startups do not make for a meaningful sample. Colonel Hervé Skenderoff, Senior Civil Servant and NATO Cooperation Officer at France’s Agence de l’Innovation de Défense (AID), offered the clearest benchmark: if DIANA’s impact is to be genuinely attributable, he told Warsight, the ideal result would be “a solution from one nation adopted by another. And I don’t think we’re there yet.” Nearly three years in, with two fully transitioned cohorts, that threshold is now about to be crossed by two start-ups: myLanguage and Kelluu.
But early signals are worth reading carefully. The challenge hub model – grouping companies by problem type rather than geography – has nonetheless produced something that goes beyond programme design: genuine, working-level collaboration across the alliance, built not on political agreement but on shared technical purpose. After three cohorts, examples are beginning to appear of those connections producing results, technologies finding their way into field trials and cross-government use, highlighting, as Crump put it, “the potential for broader cross-government and NATO-wide adoption.” At a moment when the alliance’s political cohesion is being tested more acutely than at any point in its seventy-year history, that may matter more than any individual procurement outcome.
Whether those early signals translate into contracts – Hirani-Driver’s own measure of success – remains to be seen. The structural barriers DIANA addresses have not disappeared. But the programme has begun to open a sector long defined by risk aversion and entrenched relationships. For an ecosystem that spent three decades contracting inward, even a partially opened door represents a structural change.
More subtly, it is changing who participates in defence. DIANA is building a community of innovators who have learned to navigate the system, and a growing number of officials who are learning to engage with them.
“With the war in Ukraine at Europe’s doorstep and conflicts emerging around the Gulf region, many nations are realising that defence is a whole-of-society matter,” Irimia told Warsight. For many deep tech startups, she added, the opportunity extends beyond market access: it is a chance to “bring solutions that can help end the war and move toward a more peaceful future.”
The acronym says accelerator. What it does not say is that it is also, quietly, redefining who gets to be part of the alliance’s defence.
