The deep-tech playbook you can't buy with incentives.

Start with the number that should not be possible. Between 2021 and 2025 the combined enterprise value of deep-tech startups in Estonia, Latvia and Lithuania grew 2.88 times, from €2.6 billion to €7.5 billion. No comparable European region came close. DACH — Germany, Austria, Switzerland, an economy roughly twenty times larger — managed 1.16 times. Nordic and Central European ecosystems were broadly flat. Even the United States, at 2.26 times, was outpaced by three states with a combined population smaller than metropolitan Madrid. Over the same window, deep tech went from 17.5% of all Baltic startup funding to 49.5%: nearly every second euro invested in the region now flows into hard, research-intensive technology. The figures come from the third annual Baltic Deep Tech Report, compiled by Iron Wolf Capital, the Estonian and Lithuanian national startup agencies and the law firm WALLESS on Dealroom data.

What did not produce this is the instrument most agencies reach for first. There was no national tax-holiday auction, no cash-grant bidding war between the three capitals. The machinery is duller and more durable: a deliberate pipeline from university science to commercialisation, national ecosystem bodies — Startup Estonia and Startup Lithuania, both sitting inside government innovation agencies — that do the unglamorous convening, and specialist capital that understands hardware timelines, such as Iron Wolf Capital's €100 million second fund, launched in 2025. The pitch to a founder is talent density and capital efficiency, not the lowest effective tax rate. That matters because, as this brief has argued from Poland and the OECD's global-minimum-tax rewrite, the tax-based offer is a depreciating asset. The Baltic offer is not.

The sharpest lesson is about the wedge. The Baltics took the feature of their geography that reads as a liability — a long border with Russia — and made it the specialisation. Defence, security and resilience became the fastest-accelerating dimension of the ecosystem, reaching €104 million across 47 disclosed rounds in 2025 and 15.4% of all Baltic startup funding, with 271 dual-use startups mapped and the authors calling that a floor, not a ceiling. Widen the lens and the pattern holds across the 'Tough Ten' economies from the Baltic to the Black Sea, where tough-tech ventures jumped from 7% of all startups in 2019–2022 to 16% in 2023–2025, pulled by more than €150 billion in EU defence and resilience stimulus through instruments like SAFE, EDIP and the European Defence Fund. The transferable move is not 'do defence.' It is: find the one sector where your geography is a structural advantage rather than an apology, and concentrate there.

Then there is the part agencies consistently underrate: whose money it is. In 2025, 84% of Baltic deep-tech investment came from domestic or European investors, and domestic investment has grown 238% in three years. The Baltic story is now a European capital story, not a transatlantic one — an ecosystem increasingly able to fund its own Series A through C without shipping ownership and control offshore at the first serious round. The depth is real: Series C rounds took 46% of 2025 capital and Series A activity rose 79% year on year. But the base is thinning even as the top matures — pre-seed activity nearly halved — which is the quiet risk in every fast-maturing small ecosystem: it can run out of the early-stage companies that feed the later stages before anyone notices.

The caveats are worth stating plainly, because a small ecosystem flatters easily. Concentration is severe: Estonia alone accounted for 56% of Baltic deep-tech funding, and single deals swing national totals — Latvia's Aerones raised $62 million in June 2025, about three-quarters of the country's entire deep-tech haul, and Lithuania's CAST AI booked the region's only €100 million-plus round. And for all the talk of capital staying home, the biggest exits still tilt west: a European quantum champion listing in New York this month is a reminder that scale money and public markets remain deepest in the United States. The playbook works; it is also fragile, and a couple of mega-rounds can make a thin year look like a boom.

For a promotion agency anywhere else, the asset to copy is not the defence tech. It is the method: a named consortium — a specialist fund, a national agency and a professional-services firm — publishing hard, comparable data every year so the ecosystem can be managed rather than merely celebrated; a wedge vertical matched honestly to geography; and a deliberate, years-long effort to build domestic Series A-to-C capital so that when the companies win, the region keeps them. None of that requires a subsidy the global minimum tax will erode. All of it requires patience that a roadshow cannot substitute for.