Europe Solved Its Defence Money Problem. Then It Built a Gate.

What Is This About?

One document sits behind this whole piece, and you should be able to check every line of it yourself. It is Review 01/2026, "EU defence policy in the spotlight", published by the European Court of Auditors on 3 September 2026, adopted by its Chamber III on 14 July, covering the period from 2020 to July 2026. It runs seventy-three pages. Our copy is the German edition, which is headed Analyse 01/2026.

Who publishes it matters. The European Court of Auditors is the European Union's external auditor. Its job is checking whether EU money achieves what it was meant to achieve. It is not a think tank and it has no policy line to sell.

What the document is matters just as much. A review is not an audit report. The Court gathers and structures evidence and identifies challenges and opportunities. It delivers no audit opinion, it issues no binding findings, and nobody is obliged to act on it. The Court also states that its audit mandate does not extend to all the bodies and instruments involved in EU defence policy. So this is the EU's auditor telling you what it can see, and where it cannot see.

And here is what triggered it. The Court last examined European defence in 2019, before the spending curve bent. Total EU-level defence allocations before 2021 came to €590 million. For 2021 to 2027 the figure is €11.3 billion, plus roughly €20.6 billion redirected from other programmes, plus up to €150 billion in loans on top. When money moves by that order of magnitude in four years, the auditor comes back to look.

The reason it matters to a founder or an investor is one sentence. The same document that tots up the money also contains the rulebook deciding which companies are allowed to sell into it. This piece reads that rulebook, which is where our work on Power Structures: Hidden Champions and Ecosystem Gatekeepers keeps pointing: the gate is usually written down somewhere, and somebody is already inside it.

The video goes live on Tuesday, September 15th, 2026

https://youtu.be/zctRV5efOGg

The Audio Podcast

The audio podcast goes live earlier the same day.

Subscribe to our podcasts here, or on Spotify and Apple Podcasts.

Why This Is Your Problem

The payoff goes at the front, because this is the part you can act on.

If you are a founder in this sector, a subsidised European defence market now exists: research grants, production subsidies, joint procurement, sovereign lending. It is real money and it is new money. But access to it is conditional, and the conditions are not about your technology. They are about your corporate structure, your cap table, your supply chain and your intellectual property.

Four questions decide it.

1. Where is your legal entity established?

2. Who has decisive influence over it?

3. What share of your component cost originates outside the eligible zone?

4. Who holds the design rights?

Answer all four and you have an addressable market. Fail to answer even one and what you have is a slide.

If you are an investor, those same four questions are a revenue-access variable that most defence-tech diligence does not currently price. The governance rights you negotiate, meaning board seats, vetoes, reserved matters and rights over intellectual property, are not only a control question between you and the founder. Under these instruments they can become a market-access question between the company and its customer.

And if you are simply sitting in Zurich, Vienna or Munich wondering whether any of this touches you: Germany, Austria and Switzerland stand on three different sides of these rules right now. One did not apply. One has no EU defence funding recorded at all. One is not on the eligibility list. Three neighbours, three positions, and almost nobody is writing about it in English.

One line of housekeeping for regular readers. This is the seventh instalment in The European Scale-Up Question, and it adds a fifth constraint to the four we have already named: growth capital, demand, talent and fragmentation. Today: access. The difference is that the other four are emergent market failures nobody wrote down. This one is written into EU legislation and summarised, instrument by instrument, in Annex II of the Court's review. It is the first constraint in the series you can look up, which means it is the first one you can plan around.

Two Things We Cannot Tell You

We cannot tell you why Germany did not apply for SAFE. There is no German government source stating a motive, and Germany borrows cheaply on its own, so skipping a loan can be perfectly rational treasury policy and say nothing about industrial policy. What we can tell you is that the procurement channel attached to that instrument exists whether or not Berlin uses it, and your company is subject to its rules either way.

And we cannot tell you this is only a European-champions story. The Court is blunt that preference rules can raise costs and slow delivery in the short run, and that in defence, cost data is restricted enough that compliance with those same rules may be practically difficult to verify. Plan for the rules. Do not plan on them being loosely enforced.

Nor has access replaced capital. The Court states at paragraph 14 that access to finance remains a significant obstacle, particularly for small and medium-sized companies. Capital has arrived at state level. At company level it is still a problem. What has changed is that a second gate now sits behind the first.

The Money, and What It Actually Buys

Before 2021, total EU-level defence allocations came to €590 million. That is the whole figure. For 2021 to 2027 it is €11.3 billion, plus €20.6 billion redirected from other programmes, plus up to €150 billion in loans under SAFE. Member-state budgets went from €262 billion in 2022 to €418 billion in 2025, heading for roughly €454 billion this year. Across the period, total grants come to €14.4 billion and total loans to €155.1 billion, of which SAFE alone accounts for €146.4 billion.

Four things in that stack are buyable-into, and they are not the same product. Work out which one is aimed at you before you build a go-to-market plan around the headline number.

The European Defence Fund, €7.3 billion for 2021 to 2027, is the research and development layer. It is the one with a mechanism aimed squarely at you: the EU

Defence Innovation Scheme, built to pull startups and SMEs into defence innovation. If you are pre-revenue deep tech, this is your entry point, not SAFE.

ASAP and EDIRPA, €500 million and €300 million, are production and joint-procurement money. Small, but they buy actual output.

EDIP, €1.5 billion for 2026 and 2027, is the integrated instrument, and the one carrying the strictest design-freedom conditions.

SAFE is the big number and the most misunderstood one in this sector. SAFE does not fund companies. It lends to member states, which then buy. It is not a funding round you can apply for. It is a wave of customer demand with entry rules attached, which for a vendor is either much better than a grant or entirely out of reach, depending on how you answer four questions.

From 2028 the Commission has proposed a European Competitiveness Fund whose resilience, defence-industry and space window would take about €125 billion of a proposed €234 billion, roughly 53% of the fund.

From €590 million to a proposed €125 billion window in one budget cycle is not an increase. It is a change of regime, and the rules of that regime were written while the sector was watching the headline number.

The Rulebook Nobody Reads

Most of these instruments contain what the Court calls clauses on the principle of European preference. In practice that means four conditions, set out in Annex II instrument by instrument. Here they are in the order they will hit you.

Condition 1. Establishment. Contractors must be established and operating in the European Union, Iceland, Liechtenstein or Norway. The European Defence Fund, SAFE and EDIP add Ukraine. SAFE alone adds one further category: third countries, on the basis of agreements with the European Union. Canada is named as the first non-EU country to have signed one.

Condition 2. Control. Participation by entities under the control of third countries is possible for every instrument, but subject to strict conditions.

Hear both halves of that sentence, because the first half is the half people drop. It is possible. This is not a ban on foreign investors. It is a conditional gate. And the test is not the ownership percentage on your cap table. It is whether somebody has decisive influence. A 10% holder with a board seat, vetoes and reserved matters may raise the question. A larger passive holder may not.

Condition 3. Component origin. Under EDIRPA, SAFE and EDIP, the cost of components originating outside that zone may not exceed 35%. The European Defence Fund and ASAP carry establishment and design conditions, but no numeric cap.

Be precise about what that 35% is. It caps the cost of non-eligible component origin in the thing being procured. It is not a cap on how much any one supplier may supply. Those two get confused constantly and they are different rules with different consequences.

Condition 4. Design rights and intellectual property. Results may not be subject to restrictions by third countries. Under EDIP, design freedom is required outright, with a carve-out for ammunition and missiles running to 2033.

Read those four together and the question has changed. It used to be whether a European company can win a contract. It is now: where is the entity established, who has decisive influence over it, where do its components originate, and who controls the design?

We have circled this before from the ownership side, in our work on Helsing. What Annex II adds is that the cap-table question now has a legal consequence attached, written into five regulations.

Germany, Austria and Switzerland: Three Positions

Germany, Austria and Switzerland against the EU's five defence funding instruments. Source: European Court of Auditors, Review 01/2026, Annexes II and III; defence spending per European Defence Agency, 2024.

Figures from the European Court of Auditors' Review 01/2026; 2024 spending from European Defence Agency data as cited by the Court.

Germany spent €90.6 billion on defence in 2024 on European Defence Agency figures, 2.1% of GDP, the largest absolute defence budget in the European Union. It activated the national escape clause for additional fiscal room. And it did not request a single euro under SAFE. Precision matters here: nineteen member states expressed interest in SAFE in early 2026 and Germany was not one of them. Germany was not denied anything and was not excluded from anything. Germany did not apply.

Austria spent €4.9 billion in 2024, 1.0% of GDP, one of the lowest ratios in the Union. It is an EU member and therefore fully eligible for every instrument, and it also activated the national escape clause. Austria was not among the SAFE applicants, and the Court's 2021–2027 country breakdown records no EU defence funding for Austria.

Switzerland is a different case entirely. It is not on the eligibility list in Annex II for the European Defence Fund, ASAP, EDIRPA, SAFE or EDIP. Not one of the five. Liechtenstein, population roughly forty thousand, is on all five lists, because Liechtenstein is in the European Economic Area and Switzerland is not. On the Court's funding map, Switzerland is not marked as having received nothing. It simply is not there, rendered in the same neutral fill as the United Kingdom. It is not a zero in the dataset. It is outside the dataset.

It is easy to be lazy in both directions here. Switzerland is outside because the arrangements that would put it inside do not yet exist. That is not an EU ban on Switzerland, and it is not neutrality making it impossible. The Federal Council decided on 25 June 2025 to enter exploratory talks with the European Union toward a Security and Defence Partnership, stating that such a partnership is compatible with neutrality, that Switzerland would take on no legal or financial obligations through it, and that it is the prerequisite for joint procurement projects and for participation in SAFE. Bern's own position is that neutrality does not stand in the way. On 5 March 2026 the two sides signed a Joint Declaration on strengthened cooperation in foreign and security policy, together with a Framework Participation Agreement covering Swiss contributions to EU crisis-management missions. That is a real step. It is not a Security and Defence Partnership, and it does not confer SAFE eligibility.

The Council's own page gets Switzerland wrong

One further thing emerged while checking this, and it is worth recording carefully.

The Council of the European Union's own public explainer page on SAFE, last reviewed 15 June 2026, states that Ukraine and the EEA/EFTA countries may take part in common procurement on equal terms with member states, and then lists those countries as Iceland, Liechtenstein, Norway and Switzerland.

That is not what the instrument says, and you can check it in about ninety seconds.

Council Regulation (EU) 2025/1106, Article 16(3), requires contractors and subcontractors to be established, with their executive management structures, in the Union, in an EEA EFTA State or in Ukraine. The regulation defines an EEA EFTA State as a member of the European Free Trade Association which is also a member of the European Economic Area. Switzerland is in EFTA. Switzerland is not in the European Economic Area, having rejected membership by referendum in 1992. The word "Switzerland" does not appear anywhere in the regulation, and the Court's own Annex II lists Iceland, Liechtenstein and Norway. Not Switzerland.

The same Council page is internally inconsistent. Its list of partners who could qualify through the Security and Defence Partnership route, the Article 17(1) route requiring an SDP and a bilateral agreement, does not include Switzerland either.

We are not making a scandal out of a web page. A summary page carries no legal force and the regulation governs, so this is an error in communication, not a change in the law. It matters anyway. A Swiss founder or investor running a fast eligibility check will land on the official explainer long before they open the Official Journal, and an assumption formed there is an expensive one to carry into a bid. This is exactly why the fourth habit in the Eligibility Test below is not "what did the website say" but "what does the instrument say".

If You Are Outside the European Union

The rules treat you very differently depending on which "outside" you are.

If you are a US investor looking at European deep tech, nothing here bars you. The lazy reading of this piece is "European defence is closed to American money", and that reading is wrong. Third-country control is possible for every one of these instruments, in the Court's own words subject to strict conditions. What matters is not the flag on the cheque and it is not your percentage. It is decisive influence: board seats, vetoes, reserved matters, rights over intellectual property. A 10% holder with a blocking right can raise the question. A larger passive holder may not. So the practical advice is not "stay out". It is to negotiate governance rights knowing they are also a market-access variable for your portfolio company, and to get that analysis done before the term sheet rather than during a tender.

If you are a Swiss company, you are in a genuinely different position from a German or Austrian one, and it has nothing to do with the quality of your engineering. Swiss-origin content sits outside the eligible zone for EDIRPA, SAFE and EDIP, so it counts against the 35% ceiling on non-eligible component cost rather than toward the unconstrained 65%. That is not the same as saying a Swiss supplier is capped at 35% of a system, which is a misreading we hear repeated. The rule is about the origin of component cost, not about any one supplier's commercial role. But it does mean Swiss content carries a structural disadvantage in bids under those three instruments that no amount of engineering quality removes.

And the obvious workaround is not a workaround. A Swiss parent with an EU subsidiary does not solve this. The subsidiary may answer the establishment question and leave the control question wide open, because a Swiss parent can itself raise third-country control. Domicile is one of four tests, not a master key.

If you are in the UK or Canada, SAFE has a third-country route and it is the only one there is: a Security and Defence Partnership plus a bilateral agreement. Canada took it, with a partnership in June 2025 and an agreement in June 2026. The United Kingdom held a partnership and a negotiating mandate, looked at the terms, and declined on value-for-money grounds. Note what that was and was not: SAFE loans go to member states, and only member states borrow. What the Council authorised in September 2025 was a negotiation over whether British and Canadian companies and products may take part in procurement financed by those loans. It is not a precedent for Germany declining a loan.

The Eligibility Test: Run It Yourself

Four questions. Each one has a document behind it, somebody in your company who owns it, and a cost to fix that rises the longer you leave it.

Question 1. Where is the entity established?

Document: your commercial register extract, the Handelsregisterauszug or its local equivalent.

Owner: whoever holds the corporate file, usually the CFO or company secretary.

The rule: contractors must be established and operating in the EU, Iceland, Liechtenstein or Norway; the European Defence Fund, SAFE and EDIP add Ukraine.

Cost to fix: this is the easy one, and the only one of the four an outsider can check about you. If you are established in the wrong place, fixing it is a restructuring. Months, not weeks.

Question 2. Who has decisive influence?

Document: not your cap table. Your shareholders' agreement: board composition, veto rights, reserved matters, consent thresholds, information rights, and any side letter you have forgotten about.

Owner: your general counsel, or whoever negotiated the last round.

The rule: the test is decisive influence, not percentage. A 10% holder with a blocking right on budget or strategy can raise the question. A larger passive holder may not.

Cost to fix: a reserved-matters list is negotiable while you are raising. It is very expensive to renegotiate when you are not.

Question 3. What share of component cost originates outside the eligible zone?

Document: your bill of materials, costed and tagged by country of origin. Actual origin, not the supplier's billing address.

Owner: your head of hardware or supply chain.

The rule: under EDIRPA, SAFE and EDIP the ceiling is 35% of component cost.

Cost to fix: designing an alternative source path takes twelve to eighteen months. It cannot be done inside a tender window. If nobody in your company can produce that file this week, that is your answer, and it is the most common failure of the four.

Question 4. Who holds the design rights?

Documents: your IP assignments, licence terms, and any third-country export or end-use restrictions attached to a subsystem.

Owner: whoever signed your supplier contracts.

The rule: results may not be subject to third-country restrictions, and under EDIP design freedom is required outright, with a carve-out for ammunition and missiles to 2033. The question is not only who owns the IP. It is whether you can legally modify or substitute a component without a third country's permission.

Cost to fix: renegotiating a licence after the fact, if the counterparty will even reopen it.

If you cannot answer all four, you do not have a market. You have a brochure.

Put all four in one folder. That folder is either an asset that shortens every tender you enter, or the reason you discover in week three of a bid that you were never addressable. For an investor the same line reads differently: if a portfolio company cannot answer all four, its TAM is not its addressable market. There is a real difference between theoretical market size, procurement market size, and the slice a company is eligible to bid for. Most defence-tech decks show the first.

This is a commercial screening framework, not legal advice. The actual eligibility assessment depends on the specific instrument, the corporate structure and the contractual rights involved, and it requires specialist counsel. What this gives you is the list of questions to walk in with, so you are not paying a lawyer their hourly rate to explain the basics.

One Real Example

Quantum-Systems in Gilching, Bavaria, is registered at the Amtsgericht München under HRB 217004. In May 2026 Romania's defence ministry selected the company in what it describes as its first SAFE-financed procurement, so these questions are live for it rather than hypothetical. We profiled the company in Quantum Systems: Europe's Defence-Drone Decacorn-in-Waiting.

Question one can be answered from outside: the entity and its registered management sit in Bavaria, on a public filing. Question two, only partly: the company has raised from European, British and American investors, and the shareholder list of a German GmbH is a public document, but the shareholders' agreement, where decisive influence actually lives, is not. On questions three and four there is nothing public at all. Not for this company, and not for any of its peers.

We make no claim about whether Quantum-Systems meets any eligibility test. Nobody outside the company and its customer can. That is exactly the point. One of the four answers is a public record. Three of them are documents only you can produce, which is why you produce them before somebody asks.

What to Do on Monday

One. Founder or CEO, one hour. Pull the shareholders' agreement and list every consent right, veto and reserved matter, and who holds it. Not the cap table. The rights. That list is the answer to question two, and most founders have never seen it written out in one place.

Two. Hardware lead, one week. Cost the bill of materials and tag every line by country of origin. If it cannot be done in a week, that is a finding, not a delay.

Three. Before your next term sheet, not after. Price market access into the governance design. A non-European investor does not disqualify you; that is the lazy reading and it is wrong. Third-country control is possible under every one of these instruments, subject to conditions. What can bite is governance rights amounting to decisive influence. It is much cheaper to negotiate a reserved-matters list than to unwind one.

Four. Investors, put the four questions in the IC memo. Establishment, control, component origin, design rights. Treat them as revenue-access variables, not legal footnotes. If the memo says the TAM is the European defence market, the memo is wrong. The addressable market is the slice the company is eligible to bid for, and that is a smaller number you can actually estimate.

And one correction to a thing people will tell you. A Swiss company with an EU subsidiary does not solve this. The subsidiary may answer question one and leave question two wide open, because a Swiss parent can itself raise third-country control. There is no single clever move here. If someone sells you one, get a second opinion.

Predictions on Record

1. Switzerland will not have concluded a Security and Defence Partnership with the European Union by 31 December 2027, and Swiss companies will therefore remain outside SAFE eligibility. Confidence: 65%.

2. Germany will not request funds under SAFE or any successor tranche of it before 31 December 2027. Confidence: 70%.

3. The defence, space and resilience window of the European Competitiveness Fund will be agreed below €125 billion in the final 2028–2034 budget settlement. Confidence: 60%.

We are confident about the first two because both rest on stated national positions rather than on forecasting events. We are least confident about the third, because budget negotiations are not predictable from current drafts, and we would revise it quickly on evidence of a deal.

FAQ

Is the European Court of Auditors' Review 01/2026 an audit?

No. The Court publishes this as a review, not an audit report. It delivers no audit opinion and issues no binding findings, it draws primarily on publicly available information and material gathered for the purpose, and it covers 2020 to July 2026. The German edition is titled Analyse 01/2026, which is why some coverage calls it an analysis.

Who is eligible for the European Defence Fund?

Entities established and operating in the European Union, Iceland, Liechtenstein, Norway or Ukraine. Participation by entities under third-country control is possible but subject to strict conditions, and project results may not be subject to third-country restrictions.

What is SAFE, and can a startup apply for it?

SAFE, Security Action for Europe, is a €150 billion loan instrument adopted in May 2025, financed by EU bonds and lent to member states on favourable terms, repayable over up to 45 years. It finances the customer, not the company. A startup cannot apply for it. The startup and SME mechanism in this system is the EU Defence Innovation Scheme, which sits inside the European Defence Fund.

Did Germany take money from SAFE?

No. Nineteen member states expressed interest in SAFE in early 2026 and Germany was not among them. Germany appears in the Court's country breakdown with grants only, in the €0–1 billion band.

Which countries applied for SAFE funds?

Nineteen member states expressed interest in early 2026: Belgium, Bulgaria, Croatia, Cyprus, Czechia, Denmark, Estonia, Finland, France, Greece, Hungary, Italy, Latvia, Lithuania, Poland, Portugal, Romania, Slovakia and Spain.

Is Switzerland eligible for EU defence funding?

No. Switzerland is not listed as eligible for the European Defence Fund, ASAP, EDIRPA, SAFE or EDIP. SAFE allows third-country participation on the basis of an agreement with the EU, which requires a Security and Defence Partnership. Switzerland has not concluded one.

What is the 35% rule in EU defence procurement?

Under EDIRPA, SAFE and EDIP, the cost of components originating outside the European Union, Iceland, Liechtenstein, Norway or Ukraine may not exceed 35% of component costs. It caps component-origin cost. It is not a cap on any single supplier's share.

Does a US investor make a company ineligible?

Not automatically. Participation by entities under third-country control is possible under all five instruments, subject to strict conditions. The operative test is decisive influence rather than ownership percentage, and governance rights can trigger screening, mitigation measures or guarantees.

Has capital stopped being a constraint on European defence technology?

No. The Court states at paragraph 14 that access to finance remains a significant obstacle, particularly for small and medium-sized companies. Eligibility is an additional constraint, not a replacement.

What happens to these instruments after 2027?

The Commission proposes consolidating ASAP, the European Defence Fund, EDIRPA and part of EDIP into a European Competitiveness Fund for 2028–2034, whose resilience, defence-industry and space window would take roughly €125 billion of a proposed €234 billion.

Sources

• European Court of Auditors, Review 01/2026: EU defence policy in the spotlight, adopted 14 July 2026, published 3 September 2026. Available at eca.europa.eu. Figures cited above are drawn from paragraphs 10, 14, 18, 32–42, 58, 59, 62, 66–71 and 78, and from Annexes II and III.

• Swiss Federal Council, decision of 25 June 2025 on exploratory talks toward a security and defence partnership with the EU.

• European External Action Service, EU and Switzerland sign joint declaration on strengthened cooperation on foreign and security policy, 5 March 2026.

• Council Regulation (EU) 2025/1106 (SAFE) — Article 16(3) on establishment and control, Article 16(10) on the 35% component-origin ceiling, Article 16(11) on design authority, Article 17(1) on the third-country route.

• Council of the European Union, SAFE policy page, last reviewed 15 June 2026, and Security and Defence Partnerships, last reviewed 24 March 2026.

• Council of the European Union, negotiations authorised with the UK and Canada on participation in SAFE, 18 September 2025.

• Chatham House, The UK will not join the EU's new defence fund, December 2025.

• Quantum-Systems, Romania selects Quantum Systems for SAFE-financed UAS procurement, May 2026.

Related Reading from Startuprad.io

• Power Structures: Hidden Champions and Ecosystem Gatekeepers — the pillar this piece sits under

• The European Scale-Up Question — the hub this piece extends

• Helsing — Europe's $18 Billion Defence AI Bet

• Quantum Systems: Europe's Defence-Drone Decacorn-in-Waiting

• Startup News Germany, Austria, Switzerland for June 2026: The Defence Capital Supercycle

• Demand Without Deployment: Why Europe Builds Startups But Struggles to Buy From Them

• Capital Architecture: Why Europe Funds Innovation But Struggles to Finance Scale

• Talent Without Recycling: The European Scale-Up Question, Part 4

• Defense Tech Germany: KfW Research on German VCs and Dual-Use Startups

Startuprad.io works with a small number of partners who want to reach founders, operators and investors across Germany, Austria and Switzerland. If that is you, see how a partnership works.

About the Author

Jörn "Joe" Menninger is the founder and host of Startuprad.io, one of the top-listened-to startup podcasts in Germany, Austria and Switzerland. He has been covering startups and venture capital in Germany, Austria and Switzerland since 2014 and has recorded more than 1,700 interviews with founders, investors and operators, with listeners in over 100 countries. He is based in Frankfurt am Main. More about Joe.

Machine-readable identity and routing reference: startuprad.io/llm.

Created with the assistance of AI.

Automated Transcript

Joe Menninger | Startuprad.io [00:00:00]:

Germany has the largest defense budget in the European Union, €90.6 billion in 2024. The European Union put €150 billion of defense loans on the table. 19 member states applied. Germany did not. Austria did not either, and Switzerland is not even on the eligible list. Of any of the 5 instruments. If you are building defense or dual-use technology in Germany, Austria, or Switzerland, or writing checks into it, the money question in this sector has just been replaced by a different question. Not is there capital, it is are you allowed to sell into it? And that question has 4 parts.

Joe Menninger | Startuprad.io [00:00:53]:

All of them written down. This is startuprate.io. I'm Joe Menninger | Startuprad.io recording from Frankfurt am Main. The show is about startups, scale-ups, and venture capital in Germany, Austria, and Switzerland. Let me tell you exactly what is on my desk because the whole episode rests on one document and you should be able to check every word. of it yourself. It is called Review 01/2026, EU Defence Policy in the Spotlight. It was published by the European Court of Auditors on September 3rd, 2026, adopted by Chamber 3 on July 14th, and it covers the period from 2020 to July 2026.

Joe Menninger | Startuprad.io [00:01:51]:

It runs 73 pages. Our copy is the German edition, which is headed Analyse 01/2026. Who publishes it matters. The European Court of Auditors is the European Union's external auditor, the institution whose job is checking whether EU money achieves what it was meant to achieve. It is not a think tank, it is not a lobby group, and it has no policy position to sell. And what this document is matters just as much. A review, not an audit report. The court gathers and structures evidence, and it identifies challenges and opportunities.

Joe Menninger | Startuprad.io [00:02:35]:

It does not deliver an audit opinion. It does not issue binding findings, and nobody has to act on it. The court also said in its own words, That its audit mandate does not extend to all the bodies and instruments involved in the EU defense policy. So this is the EU's auditor telling you what it can see and what it cannot see. Now, what triggered it? The court has looked at the European defense in 2019. There was before everything. Before 2022, before the spending curve bent, before the EU had defence instruments worth talking about. Here is the scale of what changed in between.

Joe Menninger | Startuprad.io [00:03:26]:

Total EU-level defence allocation before 2021 came in at around €590 million. Not billion, million. For 2021 to 2027, the figure jumped to €11.3 Plus another $20.6 billion redirected, plus up to $150 billion in loans on top. Member state budgets went from $262 billion in '22 to $418 billion in 2025. When money moves by that order of magnitude in 4 years, the auditors come back to look. That is what the document is. And the reason it matters to this audience is one sentence: the same document that tots up the money also contains the rulebook deciding which companies are allowed to sell into it. Why is this your problem? I want to put the payoff at the front because you are busy and this is the part you can act on.

Joe Menninger | Startuprad.io [00:04:37]:

If you are a founder in the sector, here is This is what changed for you. A subsidized EU defense market now exists: research grants, production subsidies, joint procurement, and sovereign lending. It is real money and it's new money. But access to it is conditional, and the conditions are not about your technology. They are about your corporate structure, your cap table, your supply chain, and your intellectual property. 4 questions decide: Where is your legal entity established? What has decisive influence over it? What share of your component cost originates outside the eligible zone? We come back to that. And what holds the design rights? Answer all 4 and you have an addressable market. Fail to answer even one and what you have is a slide.

Joe Menninger | Startuprad.io [00:05:32]:

If you're an investor, the same 4 questions are revenue access variable that most defense diligence does not currently price. The governance rights you negotiate, like board seats, vetoes, reserved matters, rights over intellectual property, are not only control questions between you and the founder. Under these instruments, they can become a market access question because between the company and its customer. If you are neither But you're sitting in Zurich or Vienna or Munich wondering whether any of this touches you. Germany, Austria, and Switzerland are standing on 3 different sides of these rules right now. One did not apply, one has no funding recorded at all, and one is not on the list. 3 neighbors, 3 positions, and almost nobody's talking about it. I'll show you the evidence for all of that.

Joe Menninger | Startuprad.io [00:06:28]:

Then I come back to what you do about it on Monday morning, and I'll give you the 4 questions in a form you can run yourself. One line of housekeeping for regular listeners: this is installment number 7 of the European Scale-Up Questions and its 5th constraint on the 4 we have named— growth capital, demand, talent fragmentation. Today, access. The difference is that the other 4 are emerging market failures nobody wrote down. This one is written into EU legislation and summarized instrument by instrument in what is called Annex II of the Court's Review. It is the first constraint in this series you can actually look up, which means you can also plan around it. 2 things I cannot tell you. I cannot tell you why Germany did not apply for SAFE.

Joe Menninger | Startuprad.io [00:07:20]:

There is no German government source stating the motive, and Germany borrows cheaply on its own. Skipping a loan can be perfectly rational treasury policy and say nothing about the industrial strategy. What I can tell you is that the procurement channel attached to that instrument exists whether or not Berlin uses it, and your company is subject to its rules either way. I cannot tell you that this is only a European champion story. The court explained that these preference rules can raise costs and slow delivery in the short run, and that in defense, cost data is restricted enough that compliance with the very same rules may be practically difficult to verify. Plan for the rules. Do not plan on them being loosely enforced. The money is the only terms that matters to a company trying to sell into it.

Joe Menninger | Startuprad.io [00:08:17]:

Before '21, total EU-level defense allocation came in around €590 million. That is the whole figure. For '21 to '27, it is €11.3 billion plus €20.6 billion redirected from other programs, plus up to €150 billion in loans under SAFE, S-A-F-E, all in capital letters. Member states' budgets went from €262 billion in '22 to €418 billion in '25, heading for roughly $450 billion this year. 4 things in that stack are buyable into, and they are not the same product. The European Defense Fund, $7.3 billion for '21 to '27, is the R&D layer and is the one with a mechanism aimed squarely at you. The EU Defense Innovation Scheme, built to build startups and SMEs into defense innovation. If you are pre-revenue deep tech, this is your entry point.

Joe Menninger | Startuprad.io [00:09:20]:

Not safe. ASAP and EDIRPA, $500 million and $300 million respectively, are production and joint production money. Small, but they buy actual output. EDIP, $1.5 billion for '26 and '27, is the integrated instrument And the one carrying the strictest design freedom conditions. SAFE is the big number, $150 billion, and the one most misunderstood in this sector. SAFE does not fund companies, it lends to member states who then buy. So it is not a funding round you can apply for, it is a wave of customer demand with entry rules attached. Which for a vendor is either much better than a grand or entirely out of reach, depending on how you answer 4 questions.

Joe Menninger | Startuprad.io [00:10:16]:

And for 28, the Commission has proposed a European Competitiveness Fund whose resilience, defence, and space window could take up to €125 billion of a proposed €234 billion, roughly 53% of the fund. From €590 million to a proposed €125 billion window in one budget cycle. This is not an increase, this is a change of regime, and the rules of that regime were written while the sector was watching the headline number. This part of my solo is about the rulebook nobody reads. Now to the rules themselves. Most of these instruments contain what the court calls clauses on the principle of European preference. In practice, that is for 4 conditions set out in Annex 2 instrument by instrument. Here they are in the order they will hit you.

Joe Menninger | Startuprad.io [00:11:16]:

Condition number 1, establishment. Contractors must be established and operating in the European Union, Iceland, Liechtenstein, or Norway. The European Defense Fund SAFE and EDIP add Ukraine. SAFE alone adds more category Third countries on the basis of agreement with the EU. Canada is named as the first non-EU country to sign. Condition 2: Control. Participation by entities under the control of third countries is possible for every instrument, but subject to strict conditions. I want you to hear both halves of that sentence, because the first half is the half people miss.

Joe Menninger | Startuprad.io [00:11:57]:

This is possible. This is not a ban on foreign investors, this is a conditional gate, and the test is not the ownership percentage on your cap table, it is whether somebody has decisive influence. A 10% holder with a board seat, vetoes, and reserved matters may raise the question; a larger passive holder may not. Condition 3: Component origin. Under EDI, RPA, SAFE, and EDIP, the cost of components originating outside that zone may not exceed 35%. The European Defence Fund and ASIP carry establishment and design conditions, but not that numeric cap. And to be precise about what the 35% is, it is a cap on the costs of non-eligible Component origin in the thing being produced. It is not a cap on how much any one supplier may supply.

Joe Menninger | Startuprad.io [00:13:01]:

Those get confused constantly, and they are different rules with different consequences. Condition number 4: design rights and IP results may be subject to restriction by third countries. Under EDIP, design freedom is required outright, with a carve-out for ammunition and missiles running to 2030. Read those 4 together and you have something bigger than a procurement technicality. It used to be enough to ask, can a European company win this contract? The question now is, who has decisive influence over the company? Where does its supply chain originate? Who controls the design? And where is the entity established? Coming up after the break, the strangest table in this entire document: the map of which countries actually took this money. Germany, Austria, and Switzerland are all On it, all in 3 different places, and one of them is not on the map at all. 3 countries, 3 positions. Okay, here's the table.

Joe Menninger | Startuprad.io [00:14:08]:

Germany, €90.6 billion on defense in '24, on the defense agency figures, 2.1% of GDP, the largest absolute defense budget in the European Union. Germany activated the national escape clause for extra fiscal room. And Germany did not request a single euro under SAFE. Precision matters here. 19 member states expressed interest in SAFE in early '26. Germany is not one of them. Germany was not denied anything and was not excluded from anything. Germany simply did not apply.

Joe Menninger | Startuprad.io [00:14:44]:

In the court's country breakdown, Germany appears with brands only, only in parentheses, in the €0 to €1 billion band. Austria, €4.9 billion in '24, 1% of GDP, one of the lowest ratios in the Union. Austria is an EU member, so it has— it's fully eligible for every one of these instruments. It also activated the national escape clause. Austria was not among the SAFE applicants, and the Quartz 21-27 country breakdown records no EU defence funding for Austria. And Switzerland? Well, Switzerland is an entirely different case. Switzerland is not on the eligibility list of the Annex 2 for the European Defence Fund, for ASAP, EDI-RPA for SAFE, or EDIP None of those 5. Liechtenstein, on the other hand, is on all 5 of them, with a population of roughly 40,000 people, because Liechtenstein is in the European Economic Area.

Joe Menninger | Startuprad.io [00:15:49]:

Switzerland is not. And on the court's funding map, Switzerland is not marked as having received nothing. Switzerland simply is not on the map. It is drawn in the same neutral grey as the United Kingdom. It is not a 0 in the dataset. It is outside of the dataset. Now, this is where it is easy to be lazy in both directions. Switzerland is outside because arrangements that would put it inside do not exist yet.

Joe Menninger | Startuprad.io [00:16:16]:

That is not an EU ban on Switzerland. And it is not neutrality making it impossible either. The Swiss Federal Council decided on June 25th, 2025, to enter exploratory talks with the EU towards a security and defence partnership. It said at the time that such partnership is compatible with neutrality, that Switzerland would take on no legal or financial obligation through it, and that it is the prerequisite— sorry, hard word— for joint procurement projects and for participating in SAFE. So, Bahn's own position is that neutrality does not stand in the way. The honest sentence is that Switzerland lacks the arrangements, not the option. Then, on March 5th, 26th, the EU and Switzerland signed a Joint Defence Declaration on Strengthened Cooperation in Foreign and Security Policy, plus a Framework Participation Agreement covering Swiss contributions to EU crisis management missions. That is a real step and it deserves its due.

Joe Menninger | Startuprad.io [00:17:22]:

But it is not a security and defence partnership and it does not confer SAFE eligibility. And here's something I found while checking this which I did not expect: the Council of the European Union's own public planar page on SAFE. The plain language one, not legal text, says that Ukraine and the EEA/EFTA countries can take part in common procurement on equal terms with member states, and then lists those countries as Iceland, Liechtenstein, Norway, and Switzerland. But that is wrong, and you can check in about 90 seconds. The regulation itself Council Regulation 2025/1106, Article 16, Paragraph 3 says contractors and subcontractors must be established with their executive management in the Union, in an EEA/EFTA state, or Ukraine. And the regulation defines an EEA/EFTA state as a member of the European Free Trade Association, which is also a member of the European Economic Area. Switzerland is in the European Free Trade Association. Switzerland is not in the European Economic Area.

Joe Menninger | Startuprad.io [00:18:34]:

Swiss voters rejected that in 1992. So Switzerland does not meet the definition. The word Switzerland does not appear anywhere in the regulation, and the Court of Auditors' own Annex II lists Iceland, Liechtenstein, and Norway, not Switzerland. The explainer page contradicts itself too. Its own list of partners who could qualify through the Security and Defence Partnership route does not include Switzerland either. I'm not making a scandal out of a web page. A summary page has no legal force and the regulation governs. But if you are a Swiss founder or Swiss investor doing a quick eligibility check and the first official EU page you land on tells you your country participates on equal terms, that is the kind of thing you build a market entry assumption on.

Joe Menninger | Startuprad.io [00:19:21]:

This is exactly why the 4th question in the eligibility test is not What did the website say? It is, what does the instrument say? So, as I record this on September 26th, Swiss-origin content sits outside the eligible zone for 3 of these instruments, which means it counts against the 35% ceiling rather than towards the 65% that does not. That is a constraint written into EU law, and no amount of Swiss engineering quality lifts this. So who benefits, who loses? The direct beneficiaries are member states that both applied and already had an industrial base, and above all states where sovereign spreads are widest, because SAFE is a loan and loan's advantage scales with your borrowing cost. The people carrying the constraint are entrants everywhere and Swiss-origin suppliers specifically. What changes next? The 2028 to 2034 budget makes this bigger, not smaller. If that €125 billion window survives negotiation, eligibility stops being a footnote and becomes one of the most consequential paragraphs in European industrial policy. If you are outside the EU. Now the part for everyone listening from outside the European Union, because this show, this show has a lot of you, and the rules treat you Very differently depending on which outside you are.

Joe Menninger | Startuprad.io [00:20:52]:

If you're a US investor looking at European deep tech, nothing here bars you. I want to be unambiguous because the lazy reading of this episode is European defense is closed to American money, and this is wrong. Third country control is possible for every one of these instruments. The court's own words that it is possible. Subject to strict conditions. What matters is not the flag on the check, and it's not your percentage. It is decisive influence. Think board seats, vetoes, reserved matters, rights over IP.

Joe Menninger | Startuprad.io [00:21:29]:

A 10% holder with a blocking right can raise the question. A larger passive holder may not. So the practical advice is not stay out. It is Negotiate the government's rights knowing they are also a market access variable for your portfolio company, and get that analysis done before the term sheet, not during the tender. If you're a Swiss company— and this is the one nobody has written up properly— you are in a genuinely different position from German or Austrian ones, and it has nothing to do with the quality of your engineering. which is pretty good, by the way. Switzerland is not on the eligibility list of any of those 5 instruments. Liechtenstein, population 40,000, is on all 5 lists because Liechtenstein is in the European Economic Area and Switzerland is not.

Joe Menninger | Startuprad.io [00:22:22]:

What that means in practice is specific. Swiss-origin content sits outside the eligibility zone for EDIRPA, SAFE, and EDIP. Sorry, I'm always spelling this out so you understand it properly. It counts against the 35% ceiling on the non-eligible component cost rather than towards the unconstrained. That is not the same as saying Swiss supplier is capped at 35% of a system. That is a misreading I have heard repeated. The rule is about the origin of component costs, not about any one supplier's commercial role. But it does mean Swiss content carries a structural disadvantage in bids under those 3 instruments.

Joe Menninger | Startuprad.io [00:23:09]:

that really no amount of engineering quality can remove. And the obvious workaround is not a workaround. A Swiss parent with an EU subsidiary does not solve this. The subsidiary may answer the establishment question and leave the control question wide open, because the Swiss parent can itself raise third-country control. Domicile is one of the 4 tests, not a master key. And now, if you are in the UK or Canada, SAFE has a third-country route, and this is the only one there is: a security and defense partnership plus a bilateral agreement. Canada took it. The UK held a partnership and a mandate, looked at the terms, and declined on the value-for-money grounds.

Joe Menninger | Startuprad.io [00:23:53]:

Switzerland has neither. A joint declaration and a framework participation agreement signed on March 26th are real steps, and they are not that. The eligibility test. Run it yourself. Right, let's run it. This is the part to take notes on. 4 questions. Each one has a document behind it, nobody in your company who owns it, and a cost to fix that goes up the longer you have to leave it.

Joe Menninger | Startuprad.io [00:24:19]:

Question 1: Where is the entity established? The document is your commercial register extract, like Handelsregisterauszug or the equivalent. Contractors must be established and operated in the EU, Iceland, Liechtenstein, or Norway. The European Defense Fund, SAFe, EDIP, add Ukraine. This is the easy one and it's the only one of the 4 that an outsider can check about you. Cost to fix if you're already established in the wrong place, this is a restructuring and it's months, not weeks. Who has decisive influence? The document is not your cap table. It is your shareholders agreement. Board composition, veto rights, reserved matters, consent threshold, information rights, and any side letter you have forgotten about.

Joe Menninger | Startuprad.io [00:25:12]:

Your general counsel or whoever negotiated that last round. The test is decisive influence, not percentage. A 10% holder with a blocking right on budget or strategy can raise the question. A larger passive holder May not cost to fix. A reserved matter list is negotiable when you're raising. It is very expensive to renegotiate when you're not. What share of component cost originates outside the eligible zone? The document is your bill of material costed and tagged by country of origin, not by supplier's address By actual origin. Owner: your head of hardware or supply chain.

Joe Menninger | Startuprad.io [00:25:59]:

Under EDI-RPA, SAFE, and EDIP, the ceiling is 35% of component costs. If nobody in your company can produce that file this week, that is your answer and is not the most common failure of the 4. Cost to fix: designing an alternative source path takes 12 to 18 months. It cannot be done inside the tender window. Who holds the design rights? Question number 4. The documents are your IP assignments, your license terms, and any third-country export or end-use restrictions attached to a subsystem. Whoever signed your supplier's contract. Under EDIP, design freedom is required outright and with a carve-out for ammunition and missiles to 2033.

Joe Menninger | Startuprad.io [00:26:49]:

The question is not Only who owns the IP. It is whether you can legally modify or substitute a component without a third country's permission. Cost to fix: renegotiate the license after the fact, if your counterparty will even open it. And the line to remember: you cannot— if you cannot answer all 4, you do not have a market, you have a brochure. Put all 4 in one folder. That folder is either an asset That shortens every tender you enter, or the reason you find out in a week— in week 3 of a bid that you're never addressable. And to be clear about what this is, a commercial screening framework, not legal advice. I'm not a lawyer.

Joe Menninger | Startuprad.io [00:27:33]:

The real assessment depends on the instrument, your structure, and your contracts, and it needs counsel. What I'm giving you is the list of questions to walk in with. Um, So you're not paying a lawyer to explain the basics to you at their hourly rate. One real example, one real example to show you where the public record stops. Quantum Systems in Gielching in Bavaria, registered at the Amtsgericht Munich on May 26th. Romania's Defense Ministry selected them in what the company describes as its first safe finance procurement. So the questions are, Life for them, not hypothetical. I can answer from the outside.

Joe Menninger | Startuprad.io [00:28:17]:

Entity and registered management in Bavaria, public filing done. Partly. They have raised from European, British, and American investors, and a German GmbH shareholder list is a public document, but the shareholders agreement where decisive influence actually lives is not. Nothing public at all, not for them. Not for any of their peers. I'm making no claim about whether Quantum Systems meets any test. Nobody outside of the company and its customers can. That is exactly the point.

Joe Menninger | Startuprad.io [00:28:51]:

One of the 4 questions is on public record. 3 of them are documents only you can produce, which is why you produce them before somebody asks. What to do the next morning. Next morning, 4 things and what to do And what to do the next day. Next morning, 4 things and who does them. 1, founder or CEO, 1 hour. Pull the shareholders agreement and list every consent right, veto, and reserved matter, who— and who holds it. Not the cap table, these rights.

Joe Menninger | Startuprad.io [00:29:23]:

That list is the answer to question 2, and most founders have never seen it written out in one place. Hardware lead, 1 week. Cost your bill of materials and tag every line by country of origin. If you cannot do it in a week, that is a finding, not a delay. Before your next term sheet, not after, price market access into the governance. A non-European investor does not disqualify you. That is lazy reading and it is wrong. Third country control is possible under every of these instruments.

Joe Menninger | Startuprad.io [00:29:57]:

Subject to conditions. What can bite is governance rights amounting to decisive influence. It is much cheaper to negotiate a reserved matter list than to unwind one. Investors, put the 4 questions into the IC memo: establishment, control, component origin, design rights. Treat them as a revenue access variable, not legal footnote. If your memo says the total addressable market is the European defense market, the memo is wrong. The addressable market is the slice the portfolio company is eligible to bid for, and that is a smaller number you can actually estimate. And one correction to a thing people will tell you: a Swiss company with an EU subsidiary does not solve it.

Joe Menninger | Startuprad.io [00:30:44]:

The subsidiary may answer question 1 and leave question 2 wide open, because a Swiss parent can itself raise a third-country control. There is no single Clever move here. If somebody tells you one, get a second opinion. Our forecast on record. 3 predictions on record with confidence level so you can hold me to them. Prediction number 1: By December 1st, 2027, Switzerland will not have concluded a security and defense partnership with the EU, and Swiss companies will therefore still be outside of SAFE eligibility. Confidence: 65%. Germany will not request funds under SAFE or any successor tranche of it before December 1st, 2027.

Joe Menninger | Startuprad.io [00:31:29]:

Confidence: 70%. The defense, space, and resilience window of the European Competitiveness Fund will be agreed below €125 billion in the final 2028-2034 budget settlement. Confidence: about 60%. And Let me be explicit about where I'm confident and where I'm not. I'm reasonably confident about the first 2 because both rest on stated national positions rather than on predicting events. I'm least confident about the 3rd one because budget negotiations are not forecastable from current drafts, and I would revise it quickly on evidence of a deal. The closing. One thing to take away: Europe did not just increase its defense budget, it wrote the rules deciding which companies can sell into it.

Joe Menninger | Startuprad.io [00:32:20]:

And Germany, Austria, and Switzerland ended up on 3 different sides of those rules. When I went looking for English language coverage of that, there was almost none there. For a deep tech dual-use founder, that is the opportunity and the trap at the same sentence. The money is real and the demand is real. But your addressable market is not a function of your cap table. Your bill of materials and your IP assignments, and those are the documents, not ambitions. Build the file before somebody asks you for it. Capital was the constraint everybody was watching.

Joe Menninger | Startuprad.io [00:32:57]:

Access is the constraint many are still missing. And unlike the other 4 in this series, this one is written down, which means for once You can read the rules before you play. If this was useful, the full write-up with every source is on startupradar.io/blog, and the full questions are in there as a checklist you can hand to your own lawyer or your investment committee. Next week we are back with an interview format. Until then, I'm Joe Menninger | Startuprad.io. This was Startupradar.io, and thank you for listening.