Technology that is real.
Something a customer can test, measure or buy. Not a roadmap.
What we mean by this
Real means someone outside your company can put it under load and read a number off it. A working system, a prototype in a lab, a pilot line, a module that runs on a customer bench. It does not have to be finished and it does not have to be certified.
This is a practical requirement, not a purity test. Industrial buyers do not buy roadmaps from small companies. They buy something they can evaluate against their own process, with their own materials, on their own site. If there is nothing to evaluate, the commercial work has no object, and everything we do turns into education with no closing step at the end.
The second half is difficulty. Something can be real and still not be worth backing. If a competent engineering team inside a large company could rebuild it in six months, the buyer's cheapest option is to wait or to build. There has to be a reason that does not happen: physics, process knowledge, a materials recipe, ten years of iteration, protected IP that actually covers the useful part.
We are relaxed about maturity and strict about evidence. Early is fine. Unproven claims are not. The moment a customer's engineer tests something and the number does not match what we said, the deal is over and so is our credibility with that account.
The qualities we weigh
Something a customer can put under load
A prototype, a pilot installation, a sample run, a demonstrator that survives being handled by someone who wants it to fail. The threshold is not commercial readiness. The threshold is that a serious buyer can form an opinion from evidence rather than from your slides.
A number you can defend
Deeptech sells on a delta: more yield, less energy, tighter tolerance, longer life, lower cost per unit. We look for that number to exist, to be measured under conditions someone else would accept, and to survive a hostile question about how it was measured. A number that only holds in ideal lab conditions is a liability in an industrial conversation.
Difficulty to copy
We ask what stops a well funded incumbent from doing this in a year. Good answers sound like years of process work, a physical constraint others have not solved, or IP that covers the mechanism rather than the packaging. Weak answers sound like execution speed and a better user interface.
Honesty about what breaks
Every real system has a failure mode, a maintenance need or a condition where performance drops. Teams that name theirs first are easier to sell for, because industrial buyers are trained to look for exactly that. Teams that insist nothing breaks get found out in technical due diligence, usually at the worst moment.
How we test this
We ask what exists today that a customer could touch this month, and who outside your company has already touched it. Then we ask for the measurement: what was tested, under what conditions, by whom, and what the comparison was. We are not auditing you. We are checking whether the claim we will be repeating in the market can hold under pressure.
We also ask what happens when it fails, and what the honest limits are. If those answers come easily, technical conversations with buyers usually go well. If they come slowly, we expect trouble later, and where we have invested our own fee, that trouble is partly ours.
What a yes looks like
- A working prototype, pilot or first installationThere is a physical or running thing, not only a design and a plan.
- Performance you can prove with dataMeasured conditions, a clear baseline, and results someone else could reproduce.
- A real barrier: physics, process, IP or years of workA concrete reason a well resourced competitor cannot match this quickly.
- A clear answer to what happens when it failsKnown failure modes and limits, stated before the buyer's engineer finds them.
What a no looks like
- Slides, simulations and a roadmapNothing a buyer can evaluate, so there is nothing for a commercial process to close.
- A thin layer on top of someone else's platformReal, but copyable, so the buyer's best move is to wait.
- Claims nobody outside the company has checkedThe first serious technical review turns into a credibility problem.
- Still two years from anything a customer could touchWorth building, too early for us. Come back when there is something to test.
Why this matters for the deal
We spend real money on the commercial work before any revenue exists. That is only rational where the technology itself is the reason a customer eventually says yes.
It also sets the honest limit of what we do. We can find the right buyers, get into the conversation and run the process. We cannot make a technical advantage exist. If the product is not the reason to buy, no amount of commercial capability fixes that, and both of us lose money finding out.