UK startups raised £534.1 million in the week to 7 August, across fourteen companies backed by 44 investors. It is the kind of figure that gets quoted as a barometer of the market.
Two of those fourteen companies raised £454.3 million between them.
OLIX took £231.4 million in a Series B at a reported £2.4 billion valuation, and Volta raised £222.9 million in a growth round co-led by Andreessen Horowitz and Altimeter Capital, according to Startup Magazine’s weekly report. That is 85% of the week’s total, raised by 14% of the week’s companies.
What the other twelve actually got
Subtract those two rounds and the week looks entirely different. Twelve companies shared just under £80 million, which averages a little under £6.7 million each.
That is a perfectly healthy Series A market and a completely different story from the headline. A founder reading “£534 million raised this week” would reasonably conclude that capital is abundant. A founder reading “twelve companies averaged £6.7 million” learns something they can actually use.
Both descriptions are drawn from the same fourteen deals. The first is dominated by two outliers, and the second is what happened to everybody else.
Why weekly totals move so violently
The arithmetic here is not specific to one week. It is structural, and it explains why weekly funding figures swing by hundreds of millions without anything changing.
Venture funding is distributed with an extremely long tail. A typical seed round is a few million; a typical Series A is ten to twenty; and a late-stage infrastructure round can be several hundred. When the largest single deal is fifty times the median, the weekly total is essentially a report on whether one of those large deals happened to sign that week.
Timing makes it worse. Large rounds take months to negotiate and are announced when the paperwork completes, which has nothing to do with market conditions during that week. A round agreed in May and announced in August lands entirely in August’s figures.
So a week-on-week fall of 30% does not mean investors pulled back by 30%. It usually means one company’s lawyers finished on a Tuesday rather than the previous Friday.
Why the outliers are getting bigger
It is worth looking at what OLIX said the £231.4 million is actually for, because it explains why this particular category of round distorts the figures so badly.
The company builds AI inference hardware: the X-1, a rack-scale inference platform, and the DX-1, a decode accelerator. The stated uses of the money are manufacturing, supply-chain commitments and senior hires.
Those are not software costs. A company shipping silicon has to place orders with fabricators, secure allocation of components months in advance, and fund inventory long before a single unit is sold. The capital requirement is set by the physical supply chain, not by headcount, and it does not scale down.
That is a different business to a SaaS company raising £5 million to hire six engineers, and it is why the two sit in the same weekly total at a fifty-to-one ratio. As more of the funding market shifts toward AI hardware, data centres and energy, the gap between the largest round and the median round widens, and every aggregate figure that sums deal values gets less informative than it was.
Volta’s round points the same way. A growth round co-led by Andreessen Horowitz and Altimeter Capital is late-stage capital from crossover investors who also buy public equities. That money is priced against public market comparables rather than against the UK venture market, so its presence in a weekly UK total says more about global appetite for a category than about whether British investors are writing cheques.
The practical consequence for a founder is that the headline is describing a market they are not in. Hardware and late-stage growth rounds are drawn from a different pool of capital, on different terms, at different sizes. Averaging them with seed and Series A produces a number that describes neither.
The following week made the same point
The next report, covering 10 to 14 August, put the weekly total at £378.8 million across thirteen companies, with 23 founders and 53 investors involved. On the face of it, funding fell by £155 million in a week.
That week was led by a large defence-linked Series C. The individually named deals underneath it were much smaller: Mindgard, a Lancaster University spinout, raised £22.22 million to scale an AI security platform; Edgify raised £6.7 million for edge AI orchestration in physical retail; and Cytix raised £5 million in a Series A led by Northern Gritstone for its change risk platform.
Add the named rounds together and they come to a fraction of the total. Once again, one undisclosed large deal is doing most of the work, and once again the honest summary is that a dozen companies raised ordinary amounts of money.
Grants are in there too
One entry that week is worth pulling out because it is a different kind of money altogether. MatAnalytics received a £619,000 Innovate UK grant to develop CITRUS, a physics-informed AI tool that compresses thermomechanical and microstructure predictions for steelmaking from hours to seconds, with the work aimed at reducing energy use in reheating furnaces.
A grant is not an investment. It does not dilute the founders, it does not require a return, and it is awarded on the merit of the project rather than on a projection of enterprise value. Counting it inside a venture funding total is defensible as a measure of money reaching startups, and misleading as a measure of investor appetite.
It is also a neat illustration of where the constraint bites. A tool that cuts energy use in reheating furnaces is aimed squarely at the problem UK steelmakers describe as existential, where power can exceed a producer’s entire value added. That is public money funding a response to a public policy failure.
What to read instead
If weekly totals are unreliable, the question is what a founder or an investor should watch.
Deal count is far steadier than deal value, because it is not distorted by outliers. Fourteen companies one week and thirteen the next is a much more informative pair of numbers than £534.1 million and £378.8 million, and it says the market barely moved.
Median round size, if it were published, would be better still. So would counts split by stage, since seed and growth respond to different things and averaging them together obscures both.
The investor count is a useful secondary signal: 44 investors in one week and 53 the next indicates participation broadening slightly even as the headline figure fell by nearly a third. That is the opposite of what the totals imply.
The shape this fits
None of this is unique to weekly reporting. It is the same distortion that runs through the half-year figures, where a single £814 million round accounted for about 68% of the North West’s entire six-month total while the region was reported as the strongest performer outside London.
Venture capital data is dominated by outliers at every level of aggregation: by week, by region, and by sector. Any figure that sums deal values without reporting the distribution behind them will describe the largest deal rather than the market.
The £534.1 million was real. So was the £6.7 million average for everybody who was not OLIX or Volta. Only one of those numbers tells a founder what to expect.
About The Author
More Stories
Lucky Egg Grew Sales 515% to Top Britain’s Growth 100
One Month of Leadership Changes Hit John Lewis, NEXT, ASOS and Puma
HMRC’s Agent Registration Is Real but the Email Is Not