$150M Series A. $850M pre-money. 100% return in 12 months.

TL;DR

Hologen, Eric Schmidt’s stealth biotech-AI venture, is raising $150M at $850M pre-money. The interesting part isn’t “large medicine models”. It’s the positioning: drug development + diagnostics + investment company, wrapped into one. That’s much closer to an underwriting pitch than a regular AI product pitch. The “100% return in 12 months” line reads as capital structuring, not biology. What would convince me this is more than narrative: prospective calls pre-specified before unblinding, generalization across sites and protocols, and regulator-grade posture. Disclosure: I’m on the strategic advisory board of QuantHealth.

$150M Series A. $850M pre-money. “100% return in 12 months.”

STAT reports that Hologen, a very quiet biotech-AI venture co-founded by Eric Schmidt, is raising big on an even bigger promise.

The interesting part isn’t “large medicine models” (whatever that means in reality), it is actually the positioning: “drug development + diagnostics + investment company, wrapped into one.”

In my view, that’s actually much closer to an underwriting pitch than a regular AI product pitch.

The bet: if you can model patient heterogeneity well enough, you can spot late-stage assets where the “average patient” result looks disappointing but a real effect is hiding in the noise. Then you finance them yourself.

If that works, it reshapes the economics of late-stage development. If it doesn’t, it’s just smarter subgroup analysis with better branding...

About that “100% return in 12 months” line: I don’t read that as biology. I read it as capital structuring. Partnering, flipping, creative deal mechanics. Late-stage trials don’t compress to 12-month payback cycles just because the models got larger.

What would convince me this is more than narrative?

• Prospective calls. Pre-specified before unblinding, then measured honestly.
• Generalization. Across sites, protocols, populations, especially if imaging is a core input.
• And obviously, regulator-grade posture. Not post-hoc explanations, but a path FDA and EMA will actually accept.

Full disclosure: I’m on the strategic advisory board of QuantHealth, so I’m biased toward the idea that prospective forecasting is actually possible and even highly effective when done rigorously. That bias also sets my bar: show me a pre-specified call made before unblinding, then show the outcome.

Now, if I imagine Eric Schmidt reading this, the rebuttal writes itself: “We’re not selling a model. We’re building an engine with the talent, the data access, and the capital to run the hard prospective tests the field keeps avoiding.” That’s a fair rebuttal. Maybe even the right one.

But it raises a different question:

• If this approach works, who captures most of the value? The model builder or the asset owner?
• And does every successful “AI for trials” company eventually become an investment firm anyway?

Would love to hear your thoughts on this.

Key takeaways

• Hologen’s positioning — drug development + diagnostics + investment company, wrapped into one — is closer to underwriting than to AI product development.
• The bet: model patient heterogeneity well enough to spot late-stage assets where a real effect is hiding in the noise, then finance them yourself.
• “100% return in 12 months” is capital structuring language, not biology. Late-stage trials don’t compress to 12-month payback cycles just because the models got larger.
• What would prove this is more than narrative: prospective calls pre-specified before unblinding, generalization across sites and protocols, and regulator-grade posture.
• Disclosure: I sit on QuantHealth’s strategic advisory board, which biases me toward believing prospective forecasting is possible when done rigorously.
• The deeper question for the category: if this works, who captures the value — the model builder or the asset owner? Does every successful “AI for trials” company eventually become an investment firm?