COLUMN: AI is killing the Pre-seed playbook

VC in 2026 might be drunk on AI funding rounds, but are we forgetting how we’ll pay the bill? The famous Fred Wilson of Union Square Ventures, one of the OGs of the VC world, wrote years ago that the goal is always “the deals that can return the fund.” But run the numbers on today’s rounds and even a unicorn doesn’t clear that bar.

The reason? Pre-seed isn’t ‘early’ anymore.

Rounds that used to be $250k-750k are now $1m-3m. Pre-money valuations have drifted to $8m-15m. If you’re a founder raising a party round at a punchy valuation, that looks like a win (more on why it isn’t later).

For the funds writing those checks, the math simply doesn’t work anymore. Take an average lead check, around $1m on a $15m pre-money ($16m post-money), which buys roughly 6% of the company. By exit, after two or three more priced rounds without the capital to keep defending pro-rata, that stake typically shrinks to somewhere around 2-3%. Even a rare unicorn outcome then returns roughly $20-30m on the position.

That sounds like a lot until you factor in fund size and portfolio math: many early-stage funds today run $50-100m or more, and since most positions return nothing, the rare winners need to cover for everyone who didn’t make it. Looking good sitting on their own isn’t enough.

That’s the bar Wilson meant.¹ This math doesn’t clear it. The smaller checks stacked around that lead have it worse: with pre-seed inflation, a $100k-300k check barely moves the needle, even then.² Run that across a portfolio and you may get some headlines in the tech press, but not returns.

So why do investors keep writing these checks? Partly FOMO. No fund wants to be the one that missed the next big thing, math be damned. But there’s a subtler driver too, signal value. A splashy round makes a fund look like it’s picking winners, and that reputation is its own kind of currency, even when the math doesn’t back it up. It also mirrors a Silicon Valley instinct to mistake a big raise for a real business.

The examples have apprered on the ground. For instance, Stockholm is white-hot for tech and AI right now. The success of Lovable created a wave of hype the ecosystem is still riding.³

And that is showing up in the cap tables. In pre-seed deals here, where I’m based, I can often guess the cap table before the pitch: a streaming legend, a fintech legend, a gaming legend, a valuation set less by the product than by how many local celebrities were willing to wire money that day. Founders then use that same list as leverage to reel in a VC firm to lead the round at the same price.

Admittedly, I did the same thing raising Svea Solar’s early rounds and was even taught how to engineer it when growing up in the Valley, so I’m not about to be hypocritical and shame the behaviour! I have a plank in my own eye, to borrow a phrase. But I’ve since learned spraying checks at inflated prices isn’t a strategy. It’s a symptom of too much capital chasing a stage that was never built to absorb it.

None of this is unique to Stockholm. It’s just louder here. The same shift is happening everywhere: the earliest, highest-conviction capital is moving one stage earlier again.

Part of that is pricing: capital chasing pre-seed until it stopped making sense. But part of it is more structural. AI has made it dramatically cheaper and faster to go from idea to working product, so real value now gets created before a company has a deck, a team, or even a name. Milestones that used to take quarters, a working prototype, a few key hires, some paying pilots, now happen in weeks. Capital that wants in before that value gets priced has to move earlier too.

I noticed this shift take hold back in the Valley about 18 months ago. Now it’s starting to penetrate the Nordics, and the rest of Europe.

In practice, it looks like this: “Inception investing” is the name for it, a term American investor Ed Sim coined to describe going in before the round exists, before there’s a deck, sometimes before there’s a company. It means helping a founder sharpen an idea they already have, sometimes picking the cofounder and making the first hires, plugging them into a handful of big enterprise accounts for early pilots, then helping them raise a larger, defensible pre-seed round, one the founder and company have actually grown into, or have a clear line of sight to, not just pitched.

It’s the anti-party-round. One investor commits before anyone else even knows the company exists, instead of a dozen small checks piling in once someone else already has.

If this sounds familiar, it should. It’s the same pattern that has repeated every decade since angel investing became institutionalised. YC formalised seed in 2005. When seed got expensive, a wave of specialists formalised pre-seed around 2014. Now pre-seed inflation is pushing capital one step earlier again, into inception investing.

Sceptics will call this relabeling: Angels dressed up for the fourth time. That may be the case, but the check is now institutional and repeatable, not one rich person’s side bet, and that’s what turns a habit into a category. It’s also harder to commoditise the way pre-seed did: there’s no deck to shop and no public round to bid up. The deal lives inside a relationship before anyone else even knows it exists, for now, at least.

So far this has all been about investors. Founders have just as much riding on it. Done right, this is better for founders too, not just the fund math.

An inflated valuation isn’t free money. It’s a bill that comes due at the next round. Growth that hasn’t caught up to price shows up as a down round, or a founder gets diluted harder than they should be. I’ve lived this from the founder’s side too.

The inception model, done right, means fewer, more committed checks, and a price that lets the company grow into the next round instead of racing to justify the last one.

Founders who start at a fair price tend to go on to raise strong rounds from top-tier VCs soon after. But only if the investor writing that first check actually has the network and operating experience to deliver it. An early check and a promise aren’t enough on their own.

The label matters less than the discipline behind it. A year from now, the winners in this shift will be the ones whose founders are still building, not fundraising.

Nolan Gray is a GP at Florent Venture Partners and a former Svea Solar co-founder.