Dear all,

Dayton, Ohio. October 1905. The letter arrived on a Tuesday. The Board of Ordnance and Fortification had considered the Wright brothers' proposal. The decision: no. The machine had not been brought to "the stage of practical operation."

They had already flown. Thirty-nine minutes. Twenty-four miles. The institution had not even looked up from its desk. The brothers did not fly again for two and a half years. They turned to Europe — to France, to Britain, to anyone who would listen. The invention that would define the 20th century nearly became a French military asset.

But here is the untold part of the story. Frank Lahm, a 29-year-old US cavalry lieutenant convalescing from typhoid in Saint-Germain-en-Laye, met the Wright brothers in a French garden in the spring of 1907. By the end of that afternoon he understood two things: these men had solved powered flight, and his own institution had twice turned them away without looking.

Reassigned to Washington that August, he wrote a single memo to his superior — the general who sat on the very board that had rejected the Wrights. He wrote clearly, from expertise, to the right person, at the right moment. The contract was signed. On August 2, the U.S. Army accepted Signal Corps Airplane No. 1. The world's first military aircraft. Price: $30,000.

The technology had not changed between 1905 and 1909. What changed was one person inside a large institution who could separate signal from noise — and had the conviction to act on it.

Every large enterprise is surrounded by signals. A 12-person team that just solved a materials problem your R&D department has spent four years on. A model that makes your grid 30% more efficient. The bottleneck is never the technology. It is the Frank Lahm: the leader with enough external context to recognise what matters, enough internal credibility to champion it, and enough courage to close a contract before consensus arrives.

This is venture clienting at its most consequential. Not a programme. A human decision, made by a specific person, to commit real budget to an early-stage technology — and to take the leap before certainty, not after.

Deep Tech Momentum, together with our partners NZZ and GlassDollar, has spent the past year building a community around exactly this capability: the Guardians of European Deep Tech. A network of 300 senior innovation leaders — heads of venture clienting, chief innovation officers, R&D directors and CTOs — who are actively closing contracts with deep tech and AI startups across Europe.

To identify the 50 who do this at the highest level, GlassDollar applied its proprietary dataset of over 400,000 verified startup-corporate collaborations to the Guardian network. The result is Europe's Top 50 Deep Tech & AI Innovation Leaders — published jointly with NZZ. Not the loudest voices in the room. The ones with the most active external startup pipelines, measured not by reputation but by verified collaboration activity.

Frank Lahm was not the most powerful person in the room. He was the one close enough to the external world to know what mattered — and who acted before the institution was ready.

The 50 leaders on this list are doing the same thing across energy grids, factory floors, defence procurement, space systems, and the computing infrastructure that will define Europe's next decade. They are the demand side of the European deep tech ecosystem. Without them, the signal stays in the lab.

Every Guardian in our network is moving European deep tech and AI forward. Fifty were simply unlucky enough to be so collaborative as to fit in a single article.

On May 20–21 in Berlin, they will be in the room at DTM26. The question is whether you will be there.

With the belief that a leap for innovators is always worth taking,
Martin

Mistral AI | Paris, France

→ Mistral AI was founded in 2023 and reached unicorn status faster than most people finish onboarding.

Aleph Alpha | Heidelberg, Germany

Black Forest Labs | Freiburg, Germany

→ Black Forest Labs is based in Germany’s Black Forest region, giving one of Europe’s hottest AI startups a distinctly rural home base.

H Company | Paris, France

ElevenLabs | London, UK / Warsaw, Poland

→ ElevenLabs first went viral by letting users generate ultra-realistic celebrity-style voices in seconds.

AI can solve doctoral-level exams, yet many leading models still fail to read a simple analog clock.

That paradox captures the 2026 Stanford AI Index better than any headline. Widely seen as the most credible annual benchmark of the global AI ecosystem, the report tracks where capability, capital, talent, and geopolitical power are actually moving. Its core message: AI is accelerating fast, but not evenly and not under control.

Here are the five most essential key take-aways:

1. AI is scaling faster than any previous technology wave

Generative AI reached mass adoption in record time, surpassing historical diffusion curves of both the PC and the internet. Corporate uptake is now mainstream, with 88% of companies already deploying or testing AI.

→ This is no longer an emerging market. AI is becoming baseline infrastructure across every sector.

2. Frontier AI is powerful and still surprisingly fragile

Top-tier models now solve advanced reasoning, coding, and graduate-level benchmark tasks. Yet many still struggle with basic real-world logic, including reading analog clocks or interpreting simple visual context.

→ Capability is improving faster than reliability. For enterprises, trust and robustness remain major bottlenecks.

3. The global AI race is now a US-China duopoly

The performance gap between US and Chinese models has narrowed sharply. The US still leads in compute, hyperscalers, and capital. China is highly competitive in patents, publications, and robotics.

→ AI leadership is consolidating around two power centers. Europe risks becoming a rule-maker without industrial leverage.

4. Industry has taken over from academia

More than 90% of major new AI models now come from companies, not universities. As competition intensifies, transparency is falling: training data, compute costs, and architectures are increasingly withheld.

→ The center of gravity has shifted from open science to proprietary systems. Control now sits with a small number of private actors.

5. Productivity gains are real, but uneven

AI delivers measurable gains in structured environments such as coding, customer support, and repetitive workflows. But in tasks requiring judgment, strategy, or creativity, benefits are mixed — and sometimes negative.

→ The next phase is not about deploying AI everywhere. It is about knowing where it actually works.

One of Europe's largest industrials — €91 billion in global turnover, 400,000+ employees across 60+ countries — has teamed up with a top-tier venture studio (FT-ranked one of Europe's top five startup hubs in 2025) to co-build the next generation of healthcare companies from day one.

The thesis is commercial, not philanthropic. By 2040, the number of people living with major chronic conditions — diabetes, cancer, heart failure, dementia — is projected to rise by 39%. The studio is looking for founders to build across three pillars:

The offer: €275k pre-seed capital, hands-on venture studio support from a 60+ team of builders and operators, and direct access to a coalition of clinical partners to validate and scale. No pre-formed idea required. No team required.

The ask: deep domain expertise. Operators, clinicians, researchers, repeat founders — anyone with real conviction about where European healthcare is heading.

→ Apply to co-build here

A curated 2-hour working session inside DTM26, hosted by Start2 Group in partnership with Deep Tech Momentum, on a single question: How does Europe build more global deep tech winners?

Closed-room. 40 seats. Founders, investors, corporates, policymakers, and ecosystem builders aligning on how scaling actually happens across European markets — and turning that alignment into cross-border collaboration.

📍 Berlin | 21 May 2026 | Courage Room

Selected participants receive a sponsored DTM conference ticket and a reserved seat in the room.

→ Learn more here

How do large corporates and startups actually work together — and where does it still break down? Bpifrance, with academic support from UC Berkeley's Haas School of Business, is running a four-country study (Germany, US, France, Singapore) to find out.

Who they need: innovation or business unit leaders at companies with 5,000+ employees, and founders of scaleups and B2B startups.

Why bother: the findings will benchmark collaboration models across leading innovation markets and shape practical recommendations for stronger ecosystems. Germany is currently underrepresented — DTM readers especially welcome.

10 minutes. Open until end of April.

→ Take part here

This edition, we are giving away a copy of Muskism: A Journey into the Mind of Elon Musk by Quinn Slobodian and Ben Tarnoff. It’s one of the sharpest recent books examining the ideology, ambition, and contradictions shaping today’s tech elite.

Slobodian, a renowned historian and bestselling author, is known for decoding the forces behind globalization, markets, and political power. In Muskism, he turns his lens to Elon Musk as a symbol of a broader era: where technology, capital, geopolitics, and personal mythology increasingly collide.

→  We’re giving away one copy.  Reply with “MUSKISM” and we’ll randomly select a winner.

Last Week’s Winner: Congratulations to Sadeghi Amir on winning last week’s Ecosystem Gift—the book Reshuffle by Sangeet Paul Choudary.