Bananastand Capital

Is VC underwriting the future or overpaying for it?

It’s getting harder to tell whether venture capital is underwriting asymmetric upside or just collectively pushing past its own risk tolerance.

Post-money valuations have expanded to levels that would have felt aggressive even 24 months ago. Today, they’re often justified with a forward-looking statement about “where the market is going.” But that raises a more uncomfortable question: who exactly is supposed to buy these companies later?

Historically, venture had two clean exit paths; acquisition or IPO. Both are starting to look less certain.

On the acquisition side, the buyer universe hasn’t expanded at the same rate as valuations. There are still only a handful of companies capable of writing multi-billion dollar checks, and most of them are already building internally. If anything, AI has made large incumbents more vertically ambitious, not more acquisitive. Why buy when you believe you can build, especially when you have the distribution advantage?

That leaves IPOs.

But IPOs require something venture sometimes ignores: public market agreement. It’s one thing to mark a company up in a private round. It’s another thing entirely to convince public investors that the valuation, and the story behind it is real, durable, and priced correctly.

Right now, I think there’s a growing disconnect. Private markets are pricing in future dominance. Public markets are still asking for proof.

And sitting in the middle of this is seed.

A month ago, I invested in a “human-in-the-loop” AI company. At the time, it made sense. It bridged the gap between model capability and real-world reliability. It felt pragmatic. Necessary, even.

Today, I’m not sure.

Not because the company executed poorly but because the ground beneath it is moving faster than expected. What was a feature gap a few months ago is quickly becoming table stakes. What required human augmentation is increasingly being automated. The question isn’t whether the company is good. It’s whether the problem it solves will still exist in the same form long enough to matter.

That’s the new risk in seed- not just company failure, but problem obsolescence.

Venture used to underwrite execution risk: Can this team build? Can they sell? Can they scale?

Now it has to underwrite time.

Will the window of relevance stay open long enough for the company to achieve escape velocity? Or will the market move underneath them before they get there?

This is what makes seed investing both more dangerous and more interesting than it’s ever been.

The bar is no longer just “Is this a good idea?” It’s “Is this idea durable in a world where capability is compounding weekly?”

And that shifts what we should be looking for.

Not just good founders but fast learners. Not just big markets but evolving ones. Technical advantage is less important than adaptability. The ability to reposition, collapse timelines, and move with the curve instead of getting flattened by it.

Because the uncomfortable truth is this: a lot of companies being funded today will not die because they were wrong.

They’ll die because they were early to a version of the problem that no longer exists.

And if that’s the case, then the real question for venture isn’t whether valuations are too high.

It’s whether we’re even measuring the right kind of risk anymore.