The AI Valuation Mania Is A Discipline Test

Frontier AI valuations are not a market signal for your Series A. They are a headline, and pricing your round against a headline is how good companies end up with bad cap tables.

Key takeaways

• Frontier lab valuations reflect scarcity and hype, not comparable unit economics
• Founders should anchor pricing to revenue quality, retention and burn multiple
• A high reference valuation from the news cycle isn’t a floor for your round
• Investors who quote mega deals back to you are testing your discipline, not offering a benchmark
• Overpricing a round now creates a painful down round later

I’ve raised money as a founder at Obby, sold a company through Baluu’s exit, and now I sit on the other side of the table advising founders through VAQA on exactly these decisions. I’ve watched three hype cycles play out from inside boardrooms, and the pattern never changes: the loudest number in the market becomes the anchor everyone reaches for, whether or not it applies to them.

This piece is about why the current AI valuation environment is dangerous for ordinary founders to copy, and what to price your raise on instead.

Why frontier valuations don’t translate downward

When a lab raises capital at a valuation that implies tens of billions in enterprise value, that number is built on scarcity of compute, scarcity of talent and a handful of investors chasing a category-defining position. It has almost nothing to do with revenue multiples, churn, or gross margin. OpenAI and Anthropic are pricing access to a technology platform that may reshape entire industries; a Series A SaaS company is pricing a product with a sales cycle, a support team and a renewal rate.

Founders read the headline number and unconsciously recalibrate what “a good valuation” looks like. I see it in almost every pitch deck review I do now. A founder with genuinely strong fundamentals will ask for a multiple that would have seemed absurd three years ago, purely because the ambient noise has shifted their sense of normal.

That instinct is understandable. It’s also a trap.

What should actually set your price

Your valuation should come from your own numbers, not the market’s mood. In my experience, the inputs that matter are boring and consistent regardless of what sector is fashionable.

• Net revenue retention, and whether it’s trending up or down
• Gross margin, adjusted for the real cost of any AI compute you’re burning
• Sales efficiency, meaning how much it costs you to land a dollar of new ARR
• Burn multiple relative to net new ARR growth
• The realistic exit multiple a buyer would pay for a business like yours today

None of these get better because a lab three rungs up the food chain raised at a huge number. If anything, elevated reference points make investors more sceptical when your actual metrics don’t support the ask, because they’ve seen too many decks trying to borrow credibility from someone else’s round.

The discipline test hiding inside the hype

I call this a discipline test because that’s genuinely what it is. Every cycle produces a set of reference points that tempt founders to skip the hard work of building a defensible number.

The founders who hold their nerve and price off their own economics tend to raise cleaner rounds with investors who actually understand the business. The founders who chase the ambient valuation environment tend to raise a round that looks great for six months and then becomes a structural problem at the next raise, because the growth needed to justify that price was never realistic.

I’ve sat on both sides of that outcome. A round priced on discipline survives a down market. A round priced on vibes does not.

A quick sanity check before you set your number

Before you finalise a valuation ask, run this test. Take your last twelve months of actual performance, forecast the next twelve conservatively, and ask whether an investor with no knowledge of the current AI narrative would still find your number reasonable. If the answer changes depending on whether they’ve read this week’s tech news, your number is built on sand.

I’d also encourage founders to model two scenarios: one where AI valuations correct within eighteen months, and one where they keep climbing. If your round only makes sense in the second scenario, you’ve priced in a bet you don’t control.

Where the real risk sits

The real risk isn’t that you ask for too little. Founders rarely leave value on the table by being conservative; investors are usually happy to pay more for a company that proves itself. The real risk is asking for a number the business can’t grow into, because that number becomes the floor every future investor measures you against.

A down round doesn’t just hurt optics. It resets employee option value, spooks existing investors and makes the next raise materially harder, regardless of how well the business is actually performing.

Pricing your round on your own terms

The mania around AI valuations will pass, the way every hype cycle eventually does. What won’t pass is the cap table you signed during it.

I’d rather see a founder raise a smaller round at a defensible price than a larger one they’ll spend two years growing into. That’s not caution for its own sake; it’s how you keep control of your company through the next three rounds, not just this one.

If you’re heading into a raise and want a second opinion on where your number should actually sit, this is core to the Finance and Fundraising work I do through VAQA. Get in touch and I’ll give you a straight answer before you’re in the room with investors.

Frequently Asked Questions

Should I mention AI valuations at all when pitching investors?

I’d avoid leaning on them as a comparable. Investors will bring up the macro environment themselves if it’s relevant, and using it to justify your own price tends to invite more scrutiny rather than less.

How do I know if my valuation ask is realistic?

Compare your ask to companies with similar revenue quality and retention that have actually closed rounds in the last six months, not to headline numbers from a different category of company. A VAQA advisory session can pressure test this before you’re in front of investors.

Does this apply if my product genuinely uses frontier AI models?

Yes, arguably more so. Using AI in your product doesn’t exempt you from being priced on unit economics; it just adds another cost line, your model spend, that needs to be modelled honestly into your margins.

What’s the biggest mistake you see founders make with valuation right now?

Anchoring their ask to the most recent big number they read about, rather than to their own trailing performance. It’s an easy trap because it feels like ambition, but it usually just transfers risk onto the next round.

Can FirstMotion or VAQA help directly with fundraising materials?

VAQA is where I focus specifically on board advisory and fundraising strategy, including valuation positioning and investor narrative. FirstMotion, which I also founded, is focused on AI Search and GEO for B2B SaaS rather than fundraising, so for a raise, VAQA is the right door.

Tom Batting is a Forbes 30 Under 30 entrepreneur, founder of Obby and Baluu, and founder of FirstMotion. He advises founders and leadership teams through VAQA on board advisory, growth, go-to-market, AI implementation, operational efficiency, and finance and fundraising.