Dear deep-tech founder,
You have spent years — perhaps a decade — developing technology that works in conditions most people have not imagined yet. Your publication record is strong. Your advisor writes that you are among the top PhD students they have mentored in twenty years. You have filed patents, won grants, presented at the best conferences in your field. And now you are trying to turn this work into a company, and it feels like the rules of the game have changed completely.
They have. This letter is an attempt to be honest with you about what the new game actually looks like, based on our experience evaluating and backing deep-tech founders across neural interfaces, neuromorphic computing, quantum systems, and advanced AI.
The Most Dangerous Phrase in Deep Tech
The phrase we hear most often from founders who ultimately don't make it is some variation of: "Once we solve the technical problem, the commercial application will be obvious." This sentence, in our experience, is almost always wrong — and founders who believe it are not ready to build a company yet.
The technical problem and the commercial application are not sequential. They are simultaneous. The shape of the technical problem you are solving should be determined by the commercial application you are targeting. If you are developing a neuromorphic inference chip and you have not yet decided whether you are targeting automotive, medical, or industrial markets, you are solving the wrong problem. Different markets require different latency profiles, different power envelopes, different form factors, different certification requirements. The "general-purpose" chip that is optimized for no specific application will be inferior to a specialized competitor in every market it enters.
The best deep-tech founders we have backed understood their target customer as deeply as they understood their technical architecture — often before they had left their university lab. CortexLabs had signed evaluation agreements with two automotive OEMs before they had a prototype chip. SynapticAI had completed a clinical feasibility study before their seed round. This is not a coincidence. It is the result of founders who thought about the commercial application as part of the research program, not as an afterthought to it.
The Team Problem No One Talks About
Here is something few investors will say directly: a team of three brilliant technical founders with no commercial experience has a very low probability of building a successful deep-tech company. Not because the founders are not talented — they almost certainly are. But because building a company is not primarily a technical challenge. It is a people challenge, a sales challenge, a communication challenge, and an organizational challenge. And the skills required to navigate these challenges are completely different from the skills required to produce excellent research.
The team structure that works in deep tech is almost always a T-shape: one or two world-class technical principals who can credibly claim to be among the top researchers in their specific domain, combined with at least one commercial leader who has actually sold complex technical products to enterprise customers. Not someone who thinks they can sell. Someone who has done it, in a domain at least adjacent to yours, with reference customers who will return our call.
This is not a comfortable thing to hear if you have spent a decade building technical expertise and are proud of your research accomplishments. We are proud of them too — that is why we are considering investing. But the commercial co-founder is not optional. It is the difference between a research project and a company. We have seen too many technically brilliant teams fail to raise Series A rounds because they could not demonstrate commercial traction. Finding a commercial co-founder before you fundraise — even if it means giving up significant equity — is almost always worth it.
What "Defensible IP" Actually Means
Every deep-tech founder we meet tells us they have defensible intellectual property. Most of them are wrong. Here is how to think about what actual IP defensibility requires.
Claims vs. Publications
A publication, however prestigious, is not IP. A publication puts your innovation in the public domain and enables competitors to build on your work. A patent with specific, well-drafted claims on the novel technical approach is IP. Before you raise a seed round, you should have filed patent applications that cover the specific technical innovations you are commercializing — not vague claims on general approaches, but specific structural and methodological claims that a competitor would have to design around.
Prior Art Risk
The neuromorphic and neural interface spaces have enormous academic publication volumes. Before we invest, we conduct patent landscape analyses to assess the risk that your core claims are anticipated by prior art. Founders who have not done this analysis themselves often discover, during diligence, that their most important claims have significant prior art risk. Conducting this analysis before fundraising — ideally with a specialized IP attorney in your domain — prevents the unpleasant surprise of losing your most important claims after you have committed to investors that they exist.
Trade Secrets and Know-How
For many deep-tech companies, the most durable competitive advantage is not patent IP but organizational know-how: the accumulated expertise in how to design, characterize, optimize, and manufacture your technology. This know-how lives in your engineers' heads and in your internal processes. It cannot be patented, but it also cannot be easily replicated by reading a patent. Document it, protect it through employment agreements and trade secret policies, and build organizational depth so it does not leave when key team members do.
The Timeline Reality Check
One of the most common mistakes deep-tech founders make in their first fundraise is underestimating how long everything takes. This is not a personal failing — it is a structural consequence of the fact that deep tech development timelines are genuinely longer than most investors expect, and most founders have internalized optimistic timelines that were reasonable in an academic setting but not in a commercial one.
Here are some realistic benchmarks from our portfolio experience:
- Hardware tape-out to first samples: 6–9 months, with significant risk of respins
- Medical device feasibility study to pivotal study: 18–36 months, depending on indication
- FDA De Novo submission to clearance: 12–24 months from submission
- Automotive supplier qualification: 24–36 months from evaluation start
- Enterprise hardware procurement cycle: 12–18 months from first meeting to purchase order
When you build your financial model for a seed fundraise, anchor it to these timelines. Do not build a model that shows product revenue in month 18 if you are shipping a hardware device that requires regulatory clearance. Investors who know the space will discount your projections accordingly. Investors who do not know the space will hold you to them when you miss — and you will miss, because the timelines are what they are.
The practical implication: raise more money than you think you need, because your timelines will slip. Plan for 24 months of runway minimum at your seed stage if you are in hardware or medical devices. The companies that die in deep tech almost always die from running out of runway before hitting a fundable milestone — not from technical failure.
How to Pitch a Deep-Tech VC
Most pitch advice is written for SaaS founders. It does not apply to you. Here is what we actually want to understand from a deep-tech pitch:
What does your technology do that nothing else can do?
This is a technical question, not a market question. We want to understand the specific innovation — the architectural novelty, the performance envelope, the mechanism of action — that enables your product to do something that existing solutions cannot. If you cannot explain this at a PhD level, we are not the right investor for you. If you can, we will spend as much time as you want going deep on the science before we discuss valuation.
Who is your first customer and why will they buy?
This is a commercial question with a very specific answer expected. Not "automotive OEMs need better edge inference" but "our first commercial customer is [specific company], we have spoken with [specific title] there, they have [specific budget and procurement process], and the purchase order will happen when [specific milestone]." If you cannot be this specific, we will help you get there — but the specificity has to exist in your thinking before it can exist in your pitch.
What is your path to a defensible market position?
Deep tech companies that succeed usually do so by establishing a reference customer relationship that makes them the de facto standard in a specific application domain. What is yours? Who is the customer that, if you win them, signals to every other buyer in the market that you are the standard? How does winning that customer create the reference that enables the next ten?
The Capital Efficiency Trap
There is a tempting narrative in some parts of the venture ecosystem that deep-tech companies should try to be capital-efficient. We disagree, at least for hardware and advanced systems companies. Capital efficiency is a virtue when you are building software and the marginal cost of development is close to zero. It is a dangerous constraint when you are building hardware and every iteration costs $500,000 and six months.
We look for founders who understand the capital requirements of their specific development roadmap honestly — not founders who have optimized their pitch for capital efficiency at the expense of realistic planning. Undercapitalized deep-tech companies make worse product decisions, move more slowly, lose key engineers to well-funded competitors, and die. Raise the amount you actually need to reach a fundable milestone, explain clearly why you need it, and be honest about the risks.
Our Commitment to You
If you read this letter and recognize yourself — if you are a founder with world-class technical depth, a clear commercial application, a balanced team, and a realistic plan — we want to meet you. We will give you our honest technical assessment, our honest view of your commercial strategy, and our honest answer on whether we are the right partner for your company at this stage.
We will not waste your time with processes that go nowhere. We make investment decisions in four to six weeks from first meeting, with no committee theater. And if we invest, we are investors for the life of your company — available for technical advisory support, co-investor introductions, customer referrals, and the full range of operational challenges that deep-tech companies face on the road from laboratory to market leader.
The work you are doing matters. The technology exists that will make neuromorphic computing ubiquitous, that will give paralyzed patients their voice back, that will make AI systems a hundred times more efficient. We want to help the best teams build those futures.
With genuine respect for what you have built,
Dr. Marcus Webb
Managing Partner, Neuron Factory