🤖 #48: Vivian's Deeptech Insider: Anthropic paid $400M for a company without a public product

Hello and welcome to edition #48 of Deeptech Insider.

Exits that look accidental are rarely accidental.

A few weeks ago, Anthropic acquired Coefficient Bio — an 8-month-old startup building AI models for biological research.

From the outside, it looked fast. Almost improbably so.

No public product. No obvious traction curve. Just a short timeline and a large outcome.

And immediately, the commentary followed the usual pattern: timing, valuation, luck, stealth execution.

But that framing misses something more important.

The real question isn’t how did this happen so quickly?

It’s:

Most founders are looking at the wrong surface

When founders think about exits, they tend to focus on what they can measure:
product maturity, revenue, funding milestones, technical progress.

All of that matters. But none of it explains why some companies become obvious to an acquirer early — while others remain invisible for years.

The difference is rarely execution speed.

It’s whether the trajectory is readable before the outcome exists.

Exits are rarely initiated. They are recognised.

One of the most consistent patterns I see across technical companies is this:

Founders assume an exit begins when a buyer appears.

In reality, it usually begins much earlier — when someone external can already interpret what the company is becoming.

Not in detail. Not with certainty.

But clearly enough that the direction feels coherent.

Recognition depends on something most founders don’t design for

This is where a more uncomfortable truth shows up.

There are companies that execute extremely well — strong teams, strong technical progress, real momentum — but still remain difficult to “place” in any coherent future category early on.

Not because they lack quality.

But because their trajectory is harder to interpret from the outside.

They are building something valuable, but the signal hasn’t yet resolved into a recognisable destination in the market’s mental models.

And in those cases, even when success eventually arrives, it often takes longer or follows more complex paths — because early recognition never had a clean shape to attach itself to.

In Coefficient Bio’s case, several things aligned:

• Founders with deep credibility in computational biology

• A specialist investor base who understood the category

• A stealth build (valuable training dataset / AI models) in a space where signal travels faster than product

• A broader market shift where large players were already moving in the same direction

Anthropic had publicly stated their life sciences direction months before. Coefficient Bio's founders knew exactly which problem to sit near.

That's not luck. That's aimed legibility.

The layer most founders don’t see: you are always building interpretation

Most founders think they are building products.

In reality, they are also building signals.

Signals about what category they belong to. What problems they are moving toward. What kind of future they are implicitly aligned with.

And those signals are being read long before any formal process begins.

This is where “early relationships” gets misunderstood.

It’s not about networking.

It’s about whether your work is already sitting inside the mental models of the people who might one day matter to your trajectory.

Coefficient Bio didn’t become legible in the eight months before acquisition.

They became legible the day they chose which labs to come from, which investor to take, and which problem space to sit near.

If your work isn’t already in those mental models?

They won’t discover you later.

They’ll reinterpret you too late.

Until then,

Vivian

P.S. Was this forwarded to you? Read past issues and subscribe here.

P.S.S.S. Are we connected on LinkedIn? I’m here.

If you have gotten value out of Vivian’s Deeptech Insider, please help us grow by sharing it with your friends!