The science is rarely the thing that stalls
By the time a deep-technology project reaches the point of a serious financing conversation, the hardest work is usually already done. The chemistry works in the lab. The pilot cycles. The patent family is filed and the claims are drafted by someone who understands what has to be defended.
Then the programme stops. Not because the physics failed, but because nobody can assemble the team that has to build it, permit it, finance it and run it.
We have spent thirty years on that second half. We have restructured a EUR 75 million industrial group, deployed more than 200 distributed-energy systems across eleven countries, and taken a company through a Frankfurt IPO. The pattern is consistent enough to state plainly: in capital-intensive technology, the binding constraint is almost never the invention. It is the operator supply.
What a year of delay actually costs
A twelve-month delay in a hardware programme is not a twelve-month delay on a spreadsheet. It is a year of salaries. It is a year of interest on money already drawn. It is a year of a competitor's head start and a year in which the founding team learns nothing new about their own machine.
The cost of that year is rarely money. It is the absence of someone who has already built the thing being built. A team that has shipped a reactor, a patent portfolio, a filing regime or a distribution channel into eleven countries does not need to be taught the sequence. They arrive knowing which decisions are reversible and which ones are expensive.
That is why we co-build rather than co-invest. Co-investing is a pricing activity: you evaluate what exists, agree a number and hold. Co-building is an assembly activity. It means absorbing the pre-bankable years inside our own balance sheet, hiring the operators, filing the patents and running the pilots that produce an operating record. That record is the asset. It is what converts a novel process into a machine a lender will underwrite.
Why the market cannot supply the operators
The obvious answer is to recruit. That answer fails for a structural reason.
The skill you need is not taught in volume anywhere. It sits at the intersection of three things no single programme trains together: the hard technical literacy to argue with an engineer, the financial literacy to argue with a credit committee, and the operational literacy to run a site in a jurisdiction where the standard playbook does not apply. You cannot hire twenty of those people in a quarter. They do not exist in the market as a queue.
So you have to manufacture them.
Manufacturing the supply
EX EPIC Academy exists for exactly this constraint. Twenty-five thousand people have applied. Twenty-five are taken per cohort. Two hundred and ten alumni now work across nineteen countries, on live projects inside real ventures rather than in lecture theatres, from the first week. There is no curriculum, no grades and no exams, because none of those things survive contact with the work.
The Academy is also the one institution in the group that crosses every company. Everything else stays inside its own lane. The Academy is deliberately not a company asset; it is the shared supply line. That is what makes rapid assembly repeatable rather than lucky, and it is why the model does not depend on a founder being in the room.
If you want to understand the argument properly, the Academy states it directly at academy.epicsolutiongroup.com, including what it refuses to be.
What the assembled record looks like
Assembly has to be evidenced, not asserted. The evidence we point to first is field deployment: more than 200 distributed-energy units designed, financed and installed across eleven countries, a capital track record above EUR 160 million across four continents, and portfolio companies operating in jurisdictions with the hardest bankability tests.
Deeper in the stack, the intellectual-property layer is run as a balance sheet asset rather than legal hygiene. EX IX began as an internal patent function and became a venture in its own right: 24 patent families filed and roughly 100 more validated in the pipeline. Above it sits EX FI, regulated financial infrastructure authorised in the United Kingdom, which is the layer that decides whether a project can be structured at all.
A pilot is a statement of intent. Hardware running in eleven countries is a statement of fact.
The public-capital layer most founders never claim
One more thing belongs in the same conversation, because it changes the arithmetic of the whole programme. In the markets we work in, governments fund between 56 and 70 percent of project capital. That money is published, allocated in advance and almost entirely unclaimed. The main reason it stays unclaimed is that nobody teaches founders how it moves.
It moves in sequence. A preparation grant buys documents: the environmental and social impact assessment, the detailed engineering design, the financial and transaction advisory pack. Those are precisely the deliverables a lender's credit committee later asks for and refuses to fund itself. A capex grant buys concrete, at up to 60 percent of a project's relevant additional costs, and is contested hard enough that it should be modelled as a probability rather than a line item. Blended and concessional finance buys risk the commercial market will not price, and on one project we studied the concessional ticket was under a fifth of the total financing while its real product was the other four fifths.
The routes that matter most in Europe are the German BAFA and project-agency instruments, the EUIPO SME Fund for intellectual-property costs, and the Horizon framework for collaborative research. Read the regulation before you read the term sheet, because the regulation is what the term sheet is priced off.
Where this is headed
On 18 to 20 November, EX-AI Summit 2026 brings the operator argument, the capital-stack argument and the deep-technology record into one room. Online, Las Palmas and Bali. Details at ex-aisummit.com.
The single most useful thing we did in thirty years was stop treating assembly as an afterthought. Founders give up more equity to a co-builder than to a passive investor at the same stage, and they give up early direction. What they buy with that equity is the team and the pre-company capital at the moment those are hardest to assemble.
If your science is stuck between the lab and the first customer, that gap is the job description. If you want this working inside your organisation, write to media@exventure.co. Advisory, board seats and media appearances are open.