Inside Our Co-Investor Network: The Real Signal in Deep-Tech Funding
There is one signal that separates an investor who backs deep tech from one who keeps backing it, and it is public, checkable, and underrated.
ARTICLE HIGHLIGHTS Raising for a hard-technology company
|
The hardest part of raising for hard technology isn't the pitch. It's that the set of investors who will underwrite science risk, hardware risk and a ten-year timeline is unusually small and almost impossible to identify from outside. A logo wall doesn't distinguish a fund that does this repeatedly from one that did it once and moved on.
This piece works through one method for telling them apart, using Earth VC's own portfolio as the worked example: the financing rounds above $10M our fund took part in, and every co-investor network of us. The method is reproducible on any company's public round history, that's the transferable part. The numbers below belong to our fund's book, not to deep tech as a whole.
It is also honest about the size of the keyhole, and about which numbers in it mean something. Some counts here describe investor behaviour. Others only describe how the sample was drawn. We say which is which.
METHODOLOGY Where this data comes from, and how wide it is
|
Which intro paths are real: who has actually shown up in the same round
A syndicate has to gel, not just individually agree. The question that matters isn't only who might invest, it's who has already sat in a round with whom, because winning one member is what makes an intro to the next one real rather than a favour. From outside, that is a black box: two funds both tagged "deep tech" may have deployed together repeatedly or never once.
"Winning one member is what makes an intro to the next one real rather than a favour."
Co-investment history is the closest observable proxy. Below is our fund's portfolio used as a worked sample, the same analysis runs on any company whose round history is public.
From our book, Alumni Ventures overlaps at five of our companies across five rounds, from 2024 to 2026. Nucleation Capital overlap at four rounds, SEEDS Capital and Y Combinator each overlap at three across three rounds.
The strongest signal is public, and hardly anyone reads it: who came back
You can't see check sizes and you can't see who led. But you can see, for any company, which investors appear in two consecutive rounds, and that is the closest public thing to proof that an investor will follow into technical risk once they know more.
We along side with Nucleation Capital, both returned to Aalo and to Blykalla. It is the only investor in this sample that followed on at both companies it overlaps with.
Valor Equity Partners appears in three consecutive Aalo rounds, April 2023 through September 2025. Lowercarbon Capital followed on at both Isometric and Supercritical. Y Combinator followed on at three separate companies. Norrsken Launcher and Armada Investment both returned to Blykalla.
Nothing about observing any of that required private information. Every one of those pairs is a date and a name in a public filing.
Run this on your own list. Take any company you're targeting an investor through. Write out its rounds in order, list the named investors in each, and mark everyone who appears in two consecutive ones. That shortlist, not the logo wall, is where your pitching time is worth spending.
"That shortlist — not the logo wall — is where your pitching time is worth spending."
Two in five of the investors here were not venture funds
Founders in hard tech get told to raise from strategics and corporate venture arms, and it is reasonable advice: in this part of the market they are a real share of the room. But a target list built only from venture funds misses a lot of who is actually in these rounds.
Across our 255 co-investors, 156 are venture funds and 99 are not.
Twenty-two are individual angels.
Eighteen are corporate venture arms and thirteen are corporates investing directly.
Eleven are accelerators or incubators.
The rest span family offices, private equity, government programs, and a handful of one-off categories.
On the question founders actually ask about strategics "Do they stay in?". The observable evidence here is small but specific. Eight corporate and CVC investors followed on at the same company:
Hitachi Ventures at Aalo,
Porsche Ventures and Amperex Technology at Group14,
SEEDS Capital at both Ampotech and BeeX,
Cap Vista and IMC Ventures at BeeX,
LCA Ventures at Treetoscope,
Anglo American at Supercritical.
Eight names is far too few to characterise how corporate capital behaves in general, and we are not going to try. What it does show is that strategic follow-on happens and is visible, so it is a question you can answer for any specific CVC on your list rather than guessing from their website.
If you're planning a local-only raise, look at where this capital was actually based
The expensive failure mode in a deep-tech raise is discovering in month four that your vertical has deal volume but no local syndicate that can assemble around it. By then the runway maths has already changed.
This data can't diagnose that for your sector. It can show how the geography looked across ourshelve co-investment network:
Of 288 appearances, 163 are investors headquartered in the Americas. Sixty are European. Thirty-two are headquartered in Asia. Nineteen more are headquartered in the Middle East; the rest are unmapped or scattered elsewhere.
The honest reading stops there. This is where the capital in these particular rounds was based, for a portfolio assembled by our fund that sources technology globally. It is consistent with the idea that deep-tech capital concentrates outside Asia; it is not evidence for it, because this data contains no rounds we were absent from.
The planning implication survives the caveat. If even an Asia-anchored deep-tech portfolio draws most of its co-investors from elsewhere, a target list built only from local venture funds is a narrow bet. Plan for global and non-VC from the first week, not from month four.
"If even an Asia-anchored deep-tech portfolio draws most of its co-investors from elsewhere, a target list built only from local venture funds is a narrow bet."
One round drew 43 investors. Here is how to read a number like that.
Syndicate size is the number founders see most and interpret most confidently, a crowded round reads as validation, a thin one as a warning. Both readings are usually unearned.
The largest here is Aalo's September 2025 round: 43 named investors besides Earth VC, in a US$100M raise. Sygaldry's April 2026 round follows at 39, in a US$139M raise.
This is the point where most analysis quietly decides what a big number means. We are not going to.
A crowded round is genuinely ambiguous. It can mean many parties independently reached conviction. It can mean the round looked safe and drew capital cheaply. It can be a function of round structure, timing, or a deal that was competitively sought rather than quietly assembled. Headcount separates none of these.
Three things would. It is worth knowing which of them you can usually get:
Check size per investor
43 investors at US$500K is a different round from 43 at US$5M
Investors came back
Whether the same investors came back for the next round at that company
Valuation
Valuation movement after the round
For Aalo, the two fields we hold say this: of the 19 investors in the July 2024 round, seven returned in September 2025, and 36 were new. Post-money moved from US$90M to US$450M between them.
Seven returning backers is one signal. Thirty-six new entrants at a five-fold valuation step is a different signal. They are in the same round, and we are not going to tell you which one won.
The habit worth taking: when you see a large syndicate, the useful question is not "how many" but "how many of them were here last time." That is findable on almost any company, and it discriminates where headcount doesn't.
"The useful question is not 'how many' but 'how many of them were here last time.'"
Chapter 2: The bigger scene
Everything above orbits Earth VC's Fund I. That's why some questions a hard-tech founder may most needs answered are absent from it. The next piece is built on a much larger dataset of deep-tech rounds backed by Asia-headquartered funds:
How does cross-border syndication between Asia and US or EU deep-tech investors actually happen?
Which regional funds actually bridge to international specialists?
Comments