The Blink Edition No. 10

The European private markets briefing you can't afford to miss. Deals, signals, and moves, decoded every Monday by SeedBlink.

For fifteen years in banking, my reflex was to reduce risk until I had enough information to act. That instinct does not transfer cleanly to building a company. Wait for certainty and the moment has usually passed.

I felt the conflict early, before SeedBlink even existed. There was no equity crowdfunding legislation in Europe at the time, and my first impulse was the banker's one: no framework, no build. We took the other path, worked out what could be done responsibly within the general legal framework, and put together a first version that was crude but functional. A few months later the draft European Crowdfunding Regulation appeared and the picture changed entirely.

Had we waited for the regulation, we would have started two years later, in a market that by then had other people in it.

The recalibration I made was to stop confusing prudence with postponement. Understand the risk, put limits around it, then decide. I didn’t unlearn banking, but I kept the discipline and gave up the need for certainty.

That distinction matters more than it sounds. Discipline means knowing what you are exposed to. Certainty means waiting until the exposure disappears, which in this business it never does.

For most of the history of dealmaking, the honest reason for a slow decision was that the information was not there yet. That excuse is becoming harder to defend.

In a survey of 1,000 senior dealmakers across 27 countries, conducted by FT Longitude for Datasite, 62% said human-only decision-making is no longer defensible in complex transactions. Half now say they regularly use AI in due diligence, the stage respondents identify as delivering the highest reported return, and 71% expect firms that ignore it to struggle within five years.

The shift is not only about expectations. In a separate McKinsey survey of 200 M&A practitioners, 40% of those using generative AI said it had shortened deal cycles by an average of 30 to 50%, with the effect most visible in due diligence, where 46% cited faster cycles and 49% reported better or additional insights. These are self-reported outcomes rather than measured ones, but they explain why the expectation has shifted.

The phrase worth sitting with is "no longer defensible". That is a statement about standards, not about tools. When analysis that once took three weeks can be compressed into three days, taking three weeks becomes a choice someone has to justify.

The same Datasite survey draws a line, though. 45% say the final decision to proceed to signing should remain entirely human. Dealmakers are using AI to compress the analytical work while protecting the commitment, which is the right instinct. Faster information does not make the future more knowable. It removes one of the reasons for standing still.

Which leaves the harder question exposed. If waiting is no longer about gathering facts, what is it about? Sometimes genuine risk that needs structuring. Sometimes conviction that has not arrived yet. And sometimes it is discomfort with deciding, disguised as diligence.

The distinction matters because the market clears with or without you. A decision deferred is still a decision, only one made by default rather than on purpose.
‍

UNTOLD, after the eleventh edition

UNTOLD wrapped its 11th edition earlier this month, and the founding team has since shared a few numbers. UNTOLD One drew more than 500,000 attendees from over 130 countries, which the team describes as its biggest edition yet based on guest feedback. The next edition, UNTOLD Star, is already selling: the first three ticket phases are sold out. The team also reports early interest from potential joint venture partners in new territories, part of the international expansion the founders outlined at the pitch.

UNTOLD ranked #3 worldwide by DJ Mag for the third year running and has stayed profitable while growing revenue nearly 28% last year, to €28.7M, with sponsors including Pepsi, Visa and Banca Transilvania adding a revenue line on top of ticketing.

The round is still open, from €1,000, with investor perks that include festival access.

Explore the round‍
‍

UK 🇬🇧

Stability AI raises $76M Series B
‍
The company behind Stable Diffusion raised $76M, bringing its total to $232M, with an unusual cap table: Universal, Sony and Warner Music all took equity alongside Electronic Arts and AMD Ventures, two years after the labels were on the other side of copyright disputes with it.

Read more

‍

Sweden 🇸🇪
‍
eComID raises $17M seed

Stockholm-based eComID raised $17M led by Systemiq Capital for a "shopping passport" that lets shoppers carry their size, fit and preferences across brands, on the premise that AI cannot personalise without context. It reaches 20 million shoppers a month across 60+ brands, with H&M Group a strategic backer since 2023.

Read more

‍

Belgium 🇧🇪
‍
MAASH secures €12.15M
‍
Brussels-based MAASH combined €5.85M in equity with Bpifrance public financing to move its mycoprotein from pilot to industrial scale, with sugar producers Tereos and Nordzucker joining as investors in a process that converts sugar into protein.

Read more
‍

AI runs on electrons
‍
Harvard's George Serafeim on the part of AI nobody photographs: a single gigawatt data centre campus consumes roughly what 830,000 US homes do. His research finds data centre returns explain about a third of daily variation in clean energy stocks, and less than a tenth for traditional energy.

Read more

‍

Can agents use a computer yet?
‍
A year ago the best computer-using model completed 42% of desktop tasks. Today's best manages 85%, above the 72% humans score. a16z on why the model has stopped being the moat, and why one operator running millions of automated tasks a month couldn't say which model executes them.

Read more

‍

Who gets rich when everyone can code

Lovable users now create a million projects a week, more in three weeks than exist on the App Store in total. Evan Armstrong on why cheaper production has never flattened a power law: in apps, the top 1% of publishers take 93% of revenue.

Read more

‍