Fireside: From Research to Market — Scaling Deep Tech

One technology, multiple possible markets — pick one deliberately

AIO's core photonic technology unlocks three distinct benefits: greater light sensitivity, better color accuracy, and smaller possible pixel sizes. Different customer segments value each differently, and each benefit implies a different product development roadmap — you cannot chase all three at once. AIO's approach: figure out early which customers value which benefit most, and focus development on that one first, validating with real customer engagement rather than developing in isolation.

They deliberately avoided the "obvious" biggest market (smartphones) first, because reaching smartphone OEMs requires product maturity and scale credibility a very early-stage company doesn't yet have. They also considered and discarded action cameras as a first market once they realized it required more feature development than their true "minimum viable product" needed. Their selection criteria for the first market: the smallest market where their core differentiator (more light) is valuable enough on its own, without needing all the other features other markets would demand — letting them focus 100% of early effort on what actually makes them different.

Build a product artifact before you have a product

Even before AIO had a working prototype, they created a realistic-looking data sheet for a hypothetical product and took it to potential customers to get honest, specific feedback — including going one level deeper to talk to their customers' customers (the end users who'd eventually buy a camera built on AIO's sensor). This let them gather structured, comparative feedback ("how important is more light to you," "what does it need to unlock this use case") without needing the actual hardware built yet, and to iterate the paper "product" across multiple market segments before committing engineering time.

The investor's litmus test: "bottom-up," not "top-down," market sizing

Gijsseling's strongest point: he sees many pitches that reason "it's a $1 trillion market, we'll capture even 10% of 10% of that, which is still huge" — technically true, but it says nothing about how the company actually gets there. He looks instead for founders who can show the concrete, "bottom-up" path: which specific customers, reachable through which specific relationships, buying which specific first product. The market being large is necessary but never sufficient.

When to raise, and from whom

Both speakers agreed fundraising and customer/product development happen in parallel, not sequentially — in deep tech, even a single meaningful experiment already costs real money, so credible fundraising has to start immediately. Gijsseling's framing of investor fit: venture capital signs you up for a specific trajectory — hyper-growth. If you want to build a smaller, capital-light, non-hyper-growth business, that's a legitimate choice, but it is fundamentally incompatible with VC money, and founders should not take VC funding expecting to run a lifestyle business with it. Different investor types also come with different implicit timelines and expectations (a "quick flip in 3 years" investor vs. one with real patience for a 10-20 year build) — founders should know their own exit horizon and pick investors whose incentives match it before taking the money, because it's very hard to renegotiate after the fact.

"Talking monkey on a pedestal" — how to decide what to build vs. outsource

A named heuristic from the conversation: given limited capital, decide what to build in-house by asking what actually creates your differentiation and IP — that's "the monkey." Everything else — the commodity infrastructure around it — is "the pedestal," and should be outsourced. The common mistake is spending early effort perfecting the pedestal (which is commodity and low-risk to outsource) instead of getting the monkey to actually talk (the genuinely hard, differentiating problem). Outsourcing the pedestal is both faster and lower-risk, since commodity work is far more likely to turn out fine when handed to a specialist than differentiated work is.

Team composition: bring in people who've done it before

Both speakers stressed that a strong technical/academic team (e.g., PhDs, postdocs) is necessary but not sufficient — investors need to believe the team, not just the technology, can execute. Concretely: hire people with real industry experience in supply chain, manufacturing, and sales — the parts of building a product a purely academic background doesn't teach. AIO's own credibility with investors was helped specifically by having people on the team who already had the relevant customer, supplier, and channel relationships, since those relationships open doors that cold outreach from a startup cannot.

Where to incorporate

Both agreed there's no strong universal preference among investors for where a company incorporates (Delaware vs. Netherlands vs. elsewhere) — some investors do have a regional mandate, but a globally-mandated investor doesn't care. AIO chose to incorporate in the Netherlands for pragmatic reasons: the specific technical talent they needed early on was concentrated there, and the local ecosystem made recruiting easier. They later opened a second office in Antwerp (Belgium) once a different phase of growth needed a different skill set concentrated there — the principle stated: let where you find talent, suppliers, and customers drive location decisions, not brand perception. A pan-European incorporation vehicle ("EU Inc") was raised as a live but unresolved policy discussion — not yet a practical option.