Pricing the Dull Deck in a Market Built for AI
Every deck on your desk this month has grown a comparables slide it cannot quite back up. Down at the bottom, in grey, sit two or three recent rounds, picked to make the ask in front of you look reasonable.
One of them is a price you remember reading about and thinking was mad. The founder read the same round and took it as the going rate.
That gap, between the round you would never have written and the one a founder treats as the floor, is what this week is about. The market has split in two, and most of the worry has gone to the wrong half.
The decks that should trouble you are not the ones with AI in the name. They are the ordinary ones, priced off a market they do not belong to.
The seed market is quoting two prices for the same kind of business this spring, and which one a founder gets has less to do with the company than with the company it claims to keep.
TechCrunch spent the season documenting the half everyone is watching: AI seed startups commanding higher valuations than anything near them, the capital and the premium pooling at the top of a market gone lopsided.
That is the consensus worry, and it is not yours. You do not write those cheques. The companies on your desk are ordinary in the way most good companies are ordinary: a real product, a slow market, a moat you can describe without a diagram.
The premium did not stay on the AI decks. It re-based the comp set your founder is anchoring to.
Pricing a round has always been an argument about comparables. The founder brings the set that flatters the ask, you bring the set that does not, and the number lands somewhere in between.
The founder is not bluffing. They are pattern-matching to the rounds they read about all spring, the same way you would. The catch is that those rounds were priced in a market you have spent a year deciding to stay out of.
So the set on the table now carries prices set by companies you would never have backed at those numbers, quoted at you as though they close the argument.
The poison is not on the AI deck you can simply decline. It is in the comparison your ordinary founder has already drawn.
Relabelling has become part of the ask. TechCrunch caught the blunt version: the same equity sold at two prices inside a single round, the label doing the work the fundamentals could not.
The softer version is on your desk. A sound company has grown an "AI" line near the top of the deck and a comparable it did not earn near the bottom.
So the work is the comp set, and refusing the wrong one is yours to do. Price the company against the businesses it actually resembles, not the ones it shares a vocabulary with.
For a dull, good company that means the dull, good evidence. Who renews, what they pay, how much it costs to keep them, who still needs this in three years. None of that improves because a deck found room to say "AI" twice.
This is the unglamorous half of the job, and it is where the returns sit. The market is busy pricing the label. Pricing the company is slower and duller, and it is most of the reason a book beats the index or does not.
There is one question that does the sorting for you.
Ask what the company looks like the morning its model provider ships the same feature as a default.
If the answer is "much the same, customers stay for X," then X is the moat, and you are pricing something real.
If the answer is a softer version of "we would have a problem," you are paying an AI price for a company carrying an AI risk, and the comp set was doing its work on you.
None of this says the AI company is overpriced. That argument is loud enough elsewhere, and it is not the one that reaches your cheque.
The one that reaches your cheque is duller: a sound, unexciting company priced off a market it does not belong to, by a comp slide nobody made you accept.
Price the moat, not the label, and the two-price market becomes someone else's argument to have. The premium is real. It is just not yours to pay.
April's cut to VCT income-tax relief, from 30% to 20%, was the kind of change that moves money without announcing where it went.
The thing to watch over the next few quarters is whether it drifts toward EIS, which kept its 30%, and what a wave of relief-chasing capital does to seed pricing at the entry point you actually write into.
Lifted Ventures and the British Business Bank are manufacturing more angels, on purpose
The Leeds network and the state-owned bank extended their partnership on 12 May to activate more angel investors outside London, particularly women; the announcement notes that only 14% of UK angels are women.
The supply of cheque-writers is now a policy programme, built on purpose and outside the capital. Your next northern co-investor was likely put in the room by an effort like this one. Worth knowing who makes your peers.
Quantum Motion raised a $160m Series C, about £120m, to put silicon qubits in the data centre
The UCL spinout's round, announced 7 May and co-led by DCVC and Kembara with the British Business Bank among the backers, funds rack-mounted quantum processors built on standard chip manufacturing.
A £120m round for a university spin-out used to be the cue to move to California. This one closed in London, with the state bank on the line. It is the patient capital the country keeps promising itself and rarely produces.
EnteroBiotix raised £19m in Glasgow to take a microbiome therapy into Phase 2b
The Glasgow company, founded by Dr James McIlroy in 2017, raised £19m led by Thairm Bio and the Scottish National Investment Bank to run a roughly 300-patient trial in irritable bowel syndrome.
Scottish life-sciences rounds share a fingerprint: a public investor anchoring it, a syndicate that has done this before, a decade-long clock everyone accepts going in. It is the slow discipline the headlines skip, and the reason the region keeps producing companies at all.
Read the lead investor, not just the lead number
Nyobolt, the Cambridge battery company, closed a $60m round this month led by Symbotic, the listed warehouse-automation firm that puts Nyobolt's cells in its robots. A customer led the round.
That is a strong signal and a complicated one. A strategic on the cap table validates the product like nothing else, and it can just as easily become the company's ceiling: the obvious buyer is now an insider, with information rights, a seat, and a price that suits them.
The dynamic travels down to your stage, where a corporate cheque is rarer but behaves the same way.
Before you welcome the strategic, ask the unglamorous question: does this investor make the company more valuable, or just more theirs?
Know an investor who mutters "but what's the actual moat" at every deck, and means it more each time? Share this issue with them.
Event18 Jun Bank of England, Monetary Policy Committee decision.
Tax31 Jul Self Assessment: second payment on account, 2025/26.
Tax5 Oct Self Assessment registration deadline, 2025/26, the return where EIS and SEIS relief is claimed.
FiscalDate TBC Autumn Budget 2026.
The Carry · thecarry.co.ukWorth a screenshot for your diary.
Replies reach the editor directly and are read and answered personally. Where would you price this differently, and what should a future issue take apart?
Two prices for one kind of company, and the half worth your worry is the quiet one: the good, ordinary business priced off a market it never joined.
Refuse the comp slide that does not fit, and most of the noise this spring stops being your problem. Plenty of honest decks are waiting for a fair hearing. Give them yours.
Until the next round.