An LOI is not an order: what actually counts as commercial proof

The signature that means nothing

Every deeptech founder I know has the same slide. It has logos on it. Under the logos, a sentence like "signed LOIs representing 4 million in pipeline". I have made that slide myself. It feels like progress. It is not, and the person sitting across from you in a funding meeting knows it is not.

Here is the uncomfortable truth about letters of intent in industrial markets: the buyer knows exactly what they signed, and it is not a commitment. It is a way to keep you warm, keep your technology off a competitor's desk, and keep their own options open, at a cost of zero euros. That is not cynicism. That is just what the document is for.

So let us be precise about what each piece of paper actually means, because founders lose years confusing them.

What an LOI actually commits the buyer to

A letter of intent commits the buyer to almost nothing. Read a few dozen and a pattern emerges: lots of "intends to evaluate", "subject to internal approval", "non-binding". The legally operative content is usually limited to confidentiality and sometimes exclusivity of discussion. Note who that protects. It is not you.

Does that make LOIs worthless? No. They have exactly two real uses. The first is internal: your champion uses the LOI to justify spending time on you, to get a pilot budget, to put your name in a planning document. The second is narrative: an LOI tells an investor that a serious company spent a lawyer's afternoon on you. That is a signal, but it is a weak one, because everyone in the room has seen startups with ten LOIs and no revenue.

The mistake is not having LOIs. The mistake is counting them as pipeline. An LOI is a measure of interest. Pipeline is a measure of process. They are not the same number.

The MOU is an LOI with a nicer font

A memorandum of understanding sounds weightier. Two parties, mutual understanding, sometimes a signing ceremony with a photo for LinkedIn. In industrial practice it commits the buyer to the same thing as an LOI: nothing enforceable about buying.

What an MOU sometimes adds is a work plan: who does what in a joint evaluation, what gets tested, by when. That has real value, because it creates calendar commitments and named engineers on the buyer's side. Calendar is a stronger signal than signature. An engineer with your test rig booked in their lab for March is worth more than a signed MOU, because their time is a budget, and budgets are real.

A purchase order is the only document that is money

Then there is the purchase order. A PO is generated inside the buyer's ERP system, tied to a cost center, approved by someone with spending authority, and it creates a legal obligation to pay you for defined goods or services at a defined price on a defined date. When a PO arrives, a procurement department has already done the work: vendor approval, budget allocation, internal sign-off.

This is why investors ask "do you have any POs?" and not "do you have any LOIs?". A PO means the machine worked. It means someone inside a large company put their name next to yours in a system that tracks consequences. Even a small PO, even a paid pilot at a fraction of your target price, is a different species of evidence than any letter.

I have watched founders argue that their LOI is "basically as strong as a PO". It is a comforting sentence and it is false, and the only people who believe it are the ones saying it.

The conversion problem

So the real question is never "how do I get more LOIs". It is: how does paper become an order? And here the pattern across companies is remarkably consistent. Paper converts when three things are true at the same time.

First, there is a defined next step with a date and a price on it, agreed before the ink dries. Not "we will evaluate and come back to you". A pilot with a scope, a start date, and a price, even a symbolic one. Money changes the conversation inside the buyer's company in a way that nothing else does, because a paying project has an owner and an owner has to report on it.

Second, the document names the path to production. The strongest pilot agreements I have seen contain a single paragraph that says: if the pilot meets criteria X, Y and Z, the parties intend to negotiate a supply agreement at roughly these volumes and this price band. Still not binding. But now the pilot exists to answer a commercial question, not a technical one, and everyone knows what success is supposed to cause.

Third, the signer has spending authority, or sits one level below someone who does. An LOI signed by an innovation scout converts at a fraction of the rate of one signed by a plant manager. This is knowable in advance. You can just ask: who signs, and what have they signed before?

The uncomfortable audit

If you have a folder of signed documents, it is worth auditing them with cold eyes. For each one, ask four questions. Is any money attached, now or at a defined future step? Is there a date by which something specific happens? Does the signer control a budget? Is the path from this document to a purchase order written down anywhere?

Four yeses means you have something close to pipeline. Zero yeses means you have a souvenir. Most collections of industrial LOIs contain both, and founders average them together into a number that feels good and means little.

The buyer-side version of this article would be short, by the way. It would say: sign LOIs freely, they cost nothing, and never let a startup mistake one for an order. Procurement people learn this in year one. Founders usually learn it in year three, at the price of a fundraise that went badly.

What to tell your investors instead

There is a version of the pipeline slide that works, and it is more honest and more impressive at the same time. It separates the three species. Paid work: POs, paid pilots, anything with an invoice. Active processes: named buyer, defined next step, date on the calendar, approval process started. Interest: LOIs, MOUs, verbal enthusiasm.

Founders resist this because the honest slide has a smaller top number. But the smaller number is believable, and the bigger one is not, and believable is the only currency that matters in that room. A deeptech company with two paid pilots and three active qualification processes has a real commercial story. A company with fourteen LOIs has a hobby.

Common questions

Is a letter of intent legally binding?

Usually no, except for specific clauses like confidentiality or exclusivity of negotiation. Courts look at the actual wording and the parties' conduct, but a standard industrial LOI does not obligate the buyer to purchase anything.

Should I charge for a pilot?

Yes, almost always. A paid pilot creates a budget owner inside the buyer's company and proves the technology survived a procurement process. The price can be symbolic. Free pilots convert to orders at dramatically lower rates because nobody inside is accountable for them.

What should be in a pilot agreement to make it convert?

Scope, dates, price, named success criteria, and a paragraph stating what happens commercially if the criteria are met: the intended volumes, price band, and the process for negotiating a supply agreement.

If your pipeline slide is built on paper that has no money, no dates and no budget holders behind it, the problem is not the slide. It is that nobody is running the conversion. That is the work we do.