NATO launched DIANA in 2023 from a simple observation: the war in Ukraine showed that many decisive technologies — commercial drones, Starlink-class systems, analytical software, image recognition — come not from traditional defence suppliers, but from civilian companies. Traditional military procurement cycles, measured in decades, could not keep up with the pace of civilian innovation.
DIANA is the structural answer to that problem. It is an accelerator with a budget in the hundreds of millions of euros, distributing capital, mentoring and access to military operators to companies built in startup logic, not in the logic of an arms contractor. The goal: shorten the path from a working prototype in a civilian lab to deployment in the armed forces — from decades to quarters.
Why Poland and why now
Poland plays a special role in the DIANA architecture. The accelerator operates through a network of regional hubs in member states; the Polish hub — co-located with FORT Kraków and AGH — serves the eastern flank of the alliance. This is not a cosmetic difference. Polish companies applying through the Polish hub work in a system that understands the operational realities of eastern Europe: hybrid threats, electronic warfare, the need for cheap scalable systems rather than luxury one-offs.
Add to that the fact that Poland is entering a phase of rapidly rising defence spending and that the Ministry of Defence is actively looking for Polish technology suppliers — and you have an environment in which DIANA is not just funding, but also one of the shorter paths to a Polish military customer.
How the program works in practice
DIANA operates in a thematic logic. NATO publishes so-called challenges — specific operational problems against which a call is opened. A company applies to a specific challenge with a specific solution. Challenges cover areas such as counter-UAS detection, satellite image analysis, critical-infrastructure cybersecurity, autonomy in the underwater and airborne domains, supply-chain resilience.
The path has two phases. Phase 1 is roughly a six-month acceleration with funding of around EUR 100,000, intensive mentoring, access to testing laboratories, and first contacts with military end users. Phase 2 — for companies that passed phase 1 successfully — comes with funding up to EUR 300,000 for development and operational validation.
For an AI company, the most important element isn't even the capital — it's the access. During the accelerator phase you meet officers from NATO forces who actually test your solution. After eight months you know whether your technology has a chance in the defence ecosystem — or, equally valuable, why it doesn't.
Which Polish AI companies have a chance
Based on previous cohorts and NATO's priority areas, you can identify the profiles that tend to pass quite precisely.
Companies working on autonomy and swarming. Drones, unmanned aerial and underwater systems, swarms, reinforcement learning for autonomous control. An area where NATO is visibly behind actors outside the alliance and is actively looking for suppliers.
Computer vision companies for ISR applications. Image analysis from satellites and drones, detection of military objects, identification of terrain changes, classification of thermal signatures. Polish companies have a natural edge here, because the NATO eastern flank generates the most real operational data.
Companies working on synthetic-content and disinformation detection. Deepfake detection, information-operations analysis, authenticity verification of image and audio. Cognitive warfare is an official NATO priority, and Polish experience in this area — both operational and research-based — is recognized.
AI companies for critical-infrastructure cybersecurity. DIANA regularly publishes challenges here, and Polish energy, port logistics and transport networks form a natural testing ground.
What NATO actually evaluates in an application
A common mistake of companies applying to DIANA is presenting technology as if it were a venture-capital pitch. That does not work. NATO does not evaluate whether you have a good financial model and a path to IPO. It evaluates four other things.
Operational value. Which specific task your technology performs better, faster, cheaper than the current solution. Not "we are changing the way the military works with data" — but "we reduce target identification time from 15 minutes to 30 seconds at the same accuracy." Concrete, metric, benchmark.
Technology readiness. NATO operates with the concept of TRL (Technology Readiness Level, scale 1–9). For DIANA phase 1, technologies at TRL 4–6 are usually accepted (validated component up to prototype in an operational environment). If you are below — it is too early. If above — target other, more deployment-focused programs.
Capacity to work with a military user. DIANA is not a fund that hands over money and waits for a report. It is an accelerator that expects close cooperation with officers testing the solution. The company must be operationally ready for feedback cycles, scope changes, information-security requirements.
Team credibility. Not in the sense of academic ranking, but in the sense of "can these people deliver." A portfolio of deployments, references, an engineering team with production experience — that weighs more than declarations.
Common mistakes by Polish applicants
Observing the first Polish DIANA cohorts and talking to companies that did not pass, I see a repeatable set of problems.
An application written in academic language. NATO does not evaluate doctorates — it evaluates products. Technical language is fine; the language of a scientific dissertation is not.
No clear definition of the operational problem. "Our technology could be useful in many defence scenarios" — that's an automatic rejection. There must be one scenario, one specific problem, one metric of improvement.
Underestimating security requirements. Working with NATO means clearance, export control, secure infrastructure, often the need to work in isolated environments. Companies that did not think about this at application stage drop out during execution.
Weak representation of defence competence in the team. It doesn't have to be a military person — but someone who knows how defence procurement works, what the lifecycle looks like, how to read requirements, helps dramatically.
How to prepare for the next call
DIANA cohorts usually open twice a year. The summer 2026 call is the target point for Polish AI companies that want to have the Polish hub as operator. A practical timeline for a company considering an application looks as follows.
Three to six months before the call — verification of fit against the published thematic priorities, selection of one challenge to apply to, consultation with the Polish hub (FORT Kraków) on whether the given technology profile is timely. This is a step companies skip, and one that can save three months of work on an application that would have no chance anyway.
Two to three months before the call — preparation of complete technical documentation, portfolio organization, obtaining letters of support from civilian customers (this matters — it shows the technology is already in use), preparation of a video presentation demonstrating the solution in action.
One month before the call — iteration on the application, review by people with NATO or MoD experience, language revision (DIANA operates in English, and language quality affects evaluation).
Who DIANA is not for
To honestly balance the perspective — DIANA is not for everyone. The companies for which this program is not a good choice are those that:
- Have purely civilian technology, with no convincing dual-use component. Trying to artificially stretch a defence narrative shows up in the panel conversation. Better to target FENG SMART or Horizon Europe.
- Are at a very early stage of technology development (TRL 1–3). DIANA does not fund basic research. Other instruments exist for that stage.
- Do not have a team operationally ready to work with a military customer. That requires time, information discipline, patience for cycles longer than VC.
Why it's worth considering even without a guarantee of success
Even a company that does not advance to DIANA phase 2 gains a few things that cannot be bought with money.
Understanding of the military customer. How operators actually think, what their priorities are, how they formulate requirements. That is knowledge which opens the door to the MoD, PGZ and other defence programs regardless of the outcome of the specific application.
Network. In the cohort you meet other companies from across NATO, military operators, potential partners for future consortia. It is a network that is hard to build from the outside in the defence sector.
External validation. The very fact of moving into a DIANA cohort is a strong market signal — and one that opens doors to investors, to strategic partnerships, to subsequent defence grants. Polish companies that have been through DIANA report an increase in the quality of investor conversations by orders of magnitude.
What to do if you're thinking of applying
The first step is free and takes an hour. You go to the DIANA site, read the current challenges, check the schedule, honestly assess whether your technology fits any of the priorities. If you paused longer on one of them — that's a signal to look deeper.
The second step — contact the Polish hub (FORT Kraków / AGH). They are responsible for supporting Polish companies through the process and can tell you realistically whether it is worth pursuing a given proposal in a given cycle. It does not replace the strategic decision, but it saves months of work on an application with no chance.
The third step — build a realistic application plan. This is not a task to do "on the side" — it requires a dedicated resource for 3–4 months. Companies that treat DIANA as something one person does on free evenings lose so much time at application stage that they might as well not have applied.