When Your Winner's Markup Costs You the Most
The note lands on a Tuesday and it is the good kind. Your best seed company has raised again, at a price several times the one you paid, and the founder has written with a warmth the quarterly updates never carry.
You reply, and you mean every word of it. The punt you took three years ago is now the line on the page that proves the whole thing works.
And then, before you have shut the laptop, the colder thought arrives. It has already done its arithmetic, and the number making you proud is a number you will never touch.
So this week is about the up-round: the most congratulated event in an angel's year, and the particular way it can cost the early backer most at the exact moment it looks like the win.
Last week was a fine week to be a founder. UKTN's tally put £480.1m into UK tech across a dozen rounds in five working days: a quantum-computing Series C, a row of AI Series As, a biotech or two, almost all of it priced and led by institutions.
Read that same list as the person who wrote the first cheque, and it is a wall of graduations. Companies leaving the stage where an angel's money does its work, and entering the one where it mostly cannot follow.
A graduation is meant to be the happy ending. For the angel who was there at the beginning, it is the moment the position changes character, and not in the direction the announcement implies.
The up-round prints a new number on your stake. On paper you are several times richer than you were on Friday.
That number is the most dangerous one in your book.
It is dangerous for three plain reasons, and a seasoned backer feels all three before naming them. You cannot bank it. There is no buyer for a seed position at someone else's Series A, so the mark is paper until an exit that is years off, if it arrives at all.
You often cannot follow it. The pro-rata that would hold your share is now a cheque at the new and much larger price, and unless you set money aside for it at the start, it is not money you have.
The notice asking for it comes with a short clock: wire your share at the new price within the window, or sign the waiver. For most early backers that is not a real choice; the money to take it up was never there to begin with.
And you may not be able to defend it. The next round re-rates everything. A flat or down round two years on marks your trophy back down, and you will have carried the markup and the markdown without a penny of either reaching your account.
Now the part the good news leaves out.
When you cannot write the pro-rata, someone else does, and they buy the slice of the upside your cheque used to carry. Super-pro-rata rights and the occasional pay-to-play clause only put that transfer in writing.
The celebrated number is the sound of you being moved down the cap table.
None of this is anyone behaving badly. In a hot priced round the founder's attention is on the new lead, the one setting the price and the terms; the early backer becomes a line to be managed rather than a party to the deal. That is simply where the gravity sits.
A markup is hardly bad news in itself, either. A marked-up book raises your next fund more easily, reassures the people whose money you deploy, and sometimes the company really is worth every turn of the new price.
The error is treating the mark as money. It is a claim on the future written in the present tense, and the spreadsheet records it as though the wire had already cleared.
That same mark then travels into your quarterly review, the one your co-investors and your own backers read, and lifts the bar you are measured against. It flatters the page and weakens your position at once, and only the first of those two facts ever shows up in a column.
The angels who come through an up-round with their holding intact did the work two years earlier. Usually they are the operator-angels, the ones who have run a cap table themselves and know what it does under pressure.
They took pro-rata and information rights in the documents while they still had the leverage to ask. They earmarked a reserve against this exact company, so a follow-on would be a decision and not a regret.
They set their line at entry, when a single page of terms was still negotiable, not in the celebration, when none of it is.
So when the next up-round note arrives, look past what the mark is worth to what you decided years ago: the reserve you set aside for this company, and whether the rights you hold still let you act on it.
If the honest replies are "nothing" and "none", the markup is not your win. It belongs to the new lead.
The up-round is still good news. It is simply someone else's, unless you arranged, a long time ago, for a share of it to be yours.
Spinout capital is organising itself. The Midlands launched Mindforge this week, a place-based fund built by eight universities; Creator Fund closed £42m for science founders.
Meanwhile the USIT guidance keeps nudging the standard university equity stake down toward 18%, from a quarter. Watch whether cheaper founder equity and patient regional money change which spinouts ever reach your desk.
SEIS funding rose 14%. Two thirds of it still went to the South East.
HMRC's statistics (21 May): SEIS companies raised £276m in 2024-25, up from £242m, while London and the South East still took 60% of EIS and 66% of SEIS.
The schemes are doing their job; the map is not. If your own dealflow only ever shows you the South East, the statistics are not describing the market. They are describing your address book.
Gnosis Health raised £1.1m to take Parkinson's-care AI out of Newcastle.
The Newcastle spinout closed £1.1m from SFC Capital, Northstar Ventures and an Innovate UK grant, for AI that helps the NHS monitor Parkinson's patients.
Angel-stage seed in a week the headlines gave to nine-figure rounds: small, regional, grant-braced. The real question is who writes the cheque after SFC, and whether they sit within 200 miles.
PhysicsX raised £225m for industrial AI, doubling its value to about £1.8bn.
The London company's Temasek-led Series C roughly doubled its valuation to about £1.8bn. PhysicsX builds AI that simulates physical engineering for manufacturers.
Note where the week's biggest cheque went: past the model wrappers, to AI with a physical product and real buyers. A state investor leading a nine-figure round into applied engineering is a vote on which AI value is real. Set it beside your own seed decks.
Run the seed cheque forward.
An entry price means nothing on its own. It is only as good as the rounds that have to follow it.
So before you write the seed cheque, run the company forward through two priced rounds of the kind the market is doing now, at the dilution they carry, and look at what is left of your stake at the end.
A seemingly fair price into a company that needs a brutal Series A it may never raise is just an expensive cheque wearing a discount.
Before the next seed cheque: if this works, who prices the next two rounds, and what is left of you when they do?
Know an investor who forwards every up-round to the group chat, and never once mentions the pro-rata they let slide? Share this issue with them.
Tax31 Jul 2026 Self Assessment: second payment on account for 2025/26.
Tax5 Oct 2026 Deadline to register for Self Assessment (new filers, 2025/26).
Tax31 Jan 2027 Self Assessment filing and balancing payment; where EIS and SEIS relief is claimed.
FiscalDate TBC Autumn Budget 2026 (HM Treasury, date not yet announced).
The Carry · thecarry.co.ukWorth a screenshot for your diary.
Replies reach the editor directly and are read and answered personally. If an up-round in your own book is sitting on a pro-rata you can't fund, what did you decide, and what should a future issue dig into?
A markup is a lovely thing to read and a poor thing to spend. When the next one lands, enjoy it, then ask what you set aside for that company while the terms were still yours. The deals you can act on beat the ones you can only admire. Go and find one.
Until the next round.