What I've learned about raising in the hardest category in venture capital and the funds that actually understand it

The founders who struggle most to raise for deep tech companies are almost never struggling because their technology isn't good enough. They're struggling because they're pitching the wrong investors. A quantum computing thesis in front of someone who quietly prefers SaaS multiples. A synthetic biology platform in front of a generalist who'll nod politely and never call back.

The mismatch is the problem. It's more common than anyone in the industry likes to admit. Before you use the list at the bottom of this page, the context below will make it significantly more useful.

Three things drove this simultaneously. AI created demand for physical infrastructure that software alone cannot satisfy. The grid has become the binding constraint on AI growth, semiconductors are crossing $1 trillion in annual sales, and US private investment in computing equipment doubled since 2023. Geopolitics rewired the risk calculus, with the number of firms actively investing in defence tech increasing 41% last year. And hardware finally got revalued. Since August 2025, tech hardware and equipment is up 16.7% and semiconductors are up 15% while software and services are down more than 24%.

"For years, the venture world treated hardware as a worse business than software. That framing has always been wrong and is totally out of date thinking today." - Celesta Capital, Deep Tech Week SF 2026

The thing most people get wrong about deep tech investing

Most generalist VCs cannot evaluate a deep tech company properly. Not because they aren't smart, but because the evaluation framework is genuinely different.

In software, you look at MRR, churn, NRR, and whether the team can sell. In deep tech, the critical signals arrive slowly and look different:

Is this actually novel science?

Or is it a better implementation of known science? The XB100, Bessemer's top private deep tech list, averages 122 patents per company, moats that protect companies even before commercialization.

What does the path from lab to manufacturing look like?

Has anyone on this team actually done it before? This question separates the funds that can add value from the ones that can only write cheques.

Who are the scientific advisors and do they actually engage?

A Nobel laureate who has met the team twice is worth less than a domain expert who is in the lab with you regularly. Investors who have been doing this long enough know the difference and will ask.

The funds on this list built their entire investment practice around asking these questions first. That is what you are actually paying for when you give up equity to one of them.

What deep tech fundraising actually looks like in 2026

A few things I've observed that the fundraising guides tend to understate:

1/ The timeline is longer and that is fine

Deep tech companies build over 5 to 10 year horizons with high capital intensity. Don't apologize for a long development timeline in front of a deep tech specialist. They are not surprised. They are evaluating whether you've planned for it.

2/ Non-dilutive capital is part of the stack, not a fallback

The US SBIR program deploys about $2.5 billion per year for early-stage deep tech R&D. The best founders use grants and institutional research partnerships to de-risk the science before they raise equity. The best deep tech VCs look for this and respect it.

3/ IP protection is foundational, not optional

Before you share detailed technical information with any investor, understand what you're sharing and with whom. Chip designs, material formulations, and manufacturing processes can lose value if they reach the wrong hands before you've built the right legal protection.

4/ Stage matching saves months

Seed deals average $4M to $6M. Series A rounds average $20M to $40M. Approaching a growth-stage fund with a pre-revenue company signals immediately that you haven't done the work.

The subsectors worth paying most attention to

Robotics is having its moment but manufacturing scale is the real question:

US robotics VC funding has tripled. Project Prometheus raised $6.2B, Figure raised $1B, Physical Intelligence raised $600M. But the most interesting companies in 2026 aren't the ones raising the largest rounds. They're the ones that have figured out how to move from prototype to production at cost.

Quantum is crossing from research to infrastructure:

Recent rounds include $148M for Origin Quantum and AUD 940M in government-backed funding for PsiQuantum. Several funds on this list have been building quantum expertise for years. They're not generalists who added quantum to their thesis last year.

Advanced materials is the most underrated category on the list:

Almost every major challenge in energy, manufacturing, aerospace, and biotech eventually comes back to materials science. The number of dedicated materials-focused funds is small relative to the importance of the category. That gap is an opportunity.

One number before you open the list

Deep tech funds yield 17% average net IRR vs 10% for traditional tech. The category is harder. The moats are more durable. The returns reflect both.

The list below covers every stage from pre-seed checks into university labs to nine-figure growth rounds, every geography from Cambridge and Helsinki to Sydney, Singapore, Tel Aviv, and Palo Alto, and every deep tech vertical you actually care about.

Read the fund descriptions carefully. Understand the thesis. Figure out which fund's expertise most closely matches the specific technical problem you're solving. Then approach the right door, not every door. Just the right one.