Your customer wants exclusivity. Excellent. Now make them pay for it.

The meeting has gone unusually well. The corporate likes the technology. The technical team wants to develop it with you. There is budget. Somebody senior has even joined the call, which in corporate-startup land is roughly the equivalent of smoke emerging from the Sistine Chapel. Then the term sheet arrives. Somewhere between confidentiality and governing law sits one innocent-looking word: exclusive. Exclusive supplier. Exclusive development partner. Exclusive rights in automotive. Exclusive rights to whatever comes out of the project. Perhaps even exclusive rights to use your technology in an entire field. It is tempting to read this as another sign that the customer really, really likes you. It is. Unfortunately, it may also mean that your first serious customer is asking you to sell part of your future company before either of you knows what that part is worth.

This is especially dangerous in deeptech because the first large customer often contributes something genuinely valuable. They have applications you cannot recreate in your lab, engineers who understand the system around your technology, qualification infrastructure, production data, money and, rather usefully, the possibility of eventually buying a lot of whatever you are building. A joint development agreement can therefore be exactly the right commercial structure. What it should not automatically become is a cheap option on every future market your technology might serve.

There is a very good recent real-world example from battery materials. In October 2025, IBU-tec announced a joint development agreement with PowerCo to develop an industrial manufacturing process for European LFP cathode material. What makes the deal interesting is not the acronym JDA. It is the way the commitments are stacked. PowerCo agreed to make technical milestone-dependent payments that could total a mid-double-digit million-euro amount over up to three years. Alongside the development agreement sits a long-term supply agreement. IBU-tec is committing to build substantial new production capacity, while PowerCo is securing the resulting production volume for at least ten years.

That is worth staring at for a moment because it is what serious exclusivity looks like. PowerCo is not saying: we are interested in your technology, please stop selling it to everybody else while we investigate. It is putting meaningful money into development, creating a path into production and making a long-term commercial commitment. IBU-tec, in return, is committing capacity. Both sides are giving something up.

That gives us the first useful rule: exclusivity is not a clause. It is a trade.

If a customer wants to restrict what you can do elsewhere, ask what economic commitment you receive in return. Development funding? Minimum annual purchases? A take-or-pay obligation? Capacity reservation fees? A licence fee? Guaranteed milestones? A defined launch program? Exclusivity without a corresponding commitment is a wonderfully asymmetric arrangement: you carry the opportunity cost and the corporate keeps the option.

This matters because early deeptech companies are particularly bad at pricing opportunity cost. Cash is visible. Markets you cannot enter are not. Imagine you have developed a coating that improves battery lifetime and your first automotive customer wants exclusive rights for “mobility”. That sounds narrower than worldwide exclusivity until somebody asks what mobility actually includes. Passenger cars? Trucks? Buses? Trains? Marine? Aviation? Micromobility? Stationary batteries installed at charging hubs? And what happens when the next application for your chemistry turns out to be worth ten times the one you are currently discussing?

The biotech industry has spent rather longer thinking about this problem. There is an excellent Andreessen Horowitz piece on the anatomy of biotech business-development deals that is useful well beyond therapeutics. It separates collaboration exclusivity, where the startup agrees not to work with others on a defined subject during the collaboration, from licence exclusivity, where a partner receives exclusive rights to use IP. It then cuts those rights further by field and territory. Their basic point is simple: even when exclusivity is justified, define the field as narrowly as you sensibly can.

Deeptech founders should steal this shamelessly.

Suppose you make a new optical sensing technology and an automotive Tier 1 wants to fund development. “Automotive” may be unnecessarily broad. Perhaps the actual collaboration concerns in-cabin driver monitoring. Give them defined rights there. Keep industrial inspection, robotics, medical imaging, aerospace and every application nobody has thought of yet. If geography matters, define that too. If the customer needs a head start rather than permanent control, time-box it. Exclusivity for eighteen months while a product launches is a very different economic proposition from exclusivity until the relevant patents expire.

The same principle appears in university technology licensing. CASRAI's recent guide to patent licensing explains why field-of-use and territory restrictions are deliberately used to prevent one licensee from swallowing every possible downstream market for a broadly useful invention. One company can hold exclusive rights in one application while the technology owner remains free to commercialise elsewhere. That architecture is particularly relevant to platform deeptech because the whole point of a platform is that the application you start with may not ultimately be the valuable one.

There is another version of exclusivity that founders should treat with similar care: IP ownership created during the development itself.

Most JDAs distinguish between background IP, meaning the technology each party already owned before the collaboration, and foreground IP, meaning what gets created during it. Background IP staying with the party that brought it in sounds obvious, but it needs to be explicit. Foreground IP is harder. If your customer contributes application knowledge while your engineers modify the core technology, who owns the improvement? “Jointly” feels pleasantly collaborative until somebody asks what joint ownership actually permits each party to do.

This is where lawyers earn their keep, and the exact answer depends on jurisdiction, so this is very much not a substitute for one. But commercially, I would be deeply uncomfortable allowing a first customer to acquire broad ownership of improvements to the underlying technology merely because those improvements happened during a customer-funded project. A cleaner structure is often to separate the platform from the application: the startup keeps improvements to its underlying technology; the customer receives defined rights to the application-specific output it paid to create. There are many possible variants, but the important thing is making the distinction before six months of engineering have made it impossible to reconstruct who contributed what.

A fascinating semiconductor example shows how sophisticated these structures can become. Everspin's disclosed joint-development arrangement with GlobalFoundries includes reciprocal IP licences, a procedure for allocating ownership of jointly developed inventions, manufacturing exclusivity for GlobalFoundries that expires after defined periods tied to qualification or completion of the relevant work, restrictions around licensing technology to named competitors, and royalties on qualifying production.

Notice what is happening there. Exclusivity has boundaries. It has duration. It is connected to a specific activity. Other rights sit around it. The commercial relationship has been engineered rather than summarised with the word “exclusive”.

And sometimes the correct answer is simply non-exclusive. Umicore and Nano One Materials explicitly described their cathode-material process collaboration as a non-exclusive joint development agreement. The two companies could combine technology and know-how without Nano One handing one partner control over the entire opportunity.

This is why I would never negotiate exclusivity as a binary yes/no question.

I would negotiate five dimensions: what, where, who, how long and in exchange for what.

What exactly is exclusive? The entire technology, a specific product, the jointly developed output, a use case, or merely supply for one named program? Where does it apply? Which geography or market? Who is restricted? Can you work with another division, another Tier 1, another OEM, a competitor? How long does the restriction survive? And what does the customer have to keep doing for the exclusivity to remain alive?

That final question is the one founders most often forget.

Exclusivity should usually have a way to decay.

If the customer promised €2 million of annual purchases and orders €200,000, why should it retain the same rights? If development stalls for twelve months, why should your market remain blocked? If the product never launches, why should exclusivity survive? If the corporate changes strategy, fires your champion and parks the project in an innovation archive somewhere, why should your startup go with it?

Tie valuable rights to valuable behaviour. Minimum purchases. Development milestones. Launch dates. Revenue thresholds. Capacity commitments. Whatever makes sense for the deal. Miss them and exclusivity narrows, expires or converts to non-exclusive rights.

There is a lovely recent photonics example of the other half of this logic. BluGlass signed a multi-phase JDA with Uviquity in 2024. The first roughly twelve-month phase came with an A$1.2 million NRE payment, with further phases expected and a follow-on master supply agreement already contemplated. In July 2026, after completing Phase I, Uviquity committed another A$1.4 million for Phase II, covering further photonics development, wafer and chip processing, packaging, integration and actual purchases of BluGlass laser products.

That is what I want founders to see when a corporate says it wants to “develop something together”. The commercial question is not merely whether the logo is impressive. It is whether the structure progressively converts technical interest into commitment.

Money is commitment. A funded engineering team is commitment. A milestone payment is commitment. A minimum purchase is commitment. A supply agreement is commitment. Building a factory is rather a lot of commitment.

A broad exclusivity clause is not commitment.

It is a request for yours.

So when the first major corporate customer asks for exclusivity, resist the urge to feel flattered for approximately thirty seconds. Then get interested.

Exclusive what?

For whom?

For how long?

And, most importantly:

what are you willing to commit in return?

If that last answer is vague, the exclusivity probably should be too.

If you're negotiating the first serious commercial structure around a deeptech collaboration and trying to work out what should be paid development, what should become supply and what you absolutely should not give away, that is exactly the kind of commercial problem SLSCRW likes getting involved in.