A contract vehicle is not a customer. Access is not demand.

Every deep tech company has a slide with a big number on it, and the number is almost always a ceiling. A contract vehicle it has been placed on. A framework agreement with a name everyone recognises. A letter of intent from a government agency. These are genuine commercial achievements and they are routinely booked as pipeline, presented to boards as demand, and used to justify hiring against revenue that has not been ordered. The practice behind this note lives at GTM consulting for deep tech and hardware.

Three announcements in eleven days, and not one purchase order.

What Amazon bought was optionality. What AutoStore won was the removal of every procurement obstacle between an Amazon site manager and an AutoStore system. That is worth a great deal and it is not the same asset as a signed quantity. The distinction matters because the two things behave differently under stress: an order survives a budget freeze, and a framework does not, because there is nothing in a framework for a freeze to cancel.

A framework agreement removes the friction from an order that has not been placed. It does not place the order.

Blue Water Autonomy, August 4, 2026. The company was selected by the Naval Oceanographic Office for autonomous open-ocean survey work under a multi-year, multiple-award indefinite-delivery, indefinite-quantity contract with a ceiling value of $40 million. The two qualifiers in that sentence are the whole story. Multiple-award means other companies were selected too. Indefinite quantity means the government has committed to a range rather than an amount. What the company won is the right to compete for individual task orders against the other awardees. The $40 million is the maximum the government may spend across every awardee over the life of the vehicle, and a vendor's share of it is determined later, by task orders that have their own competitions, their own timelines and their own funding.

Multibeam, July 29, 2026. The company signed a letter of intent with the U.S. Department of Commerce for $140 million in proposed funding from the CHIPS Research and Development Office, to accelerate work on fine-pitch advanced packaging. In the same month Commerce announced letters of intent with seven companies totalling $874 million, per the National Institute of Standards and Technology. A letter of intent is a statement that a counterparty intends to proceed subject to conditions. Multibeam's own release calls the money proposed. Alongside it the company closed a $50 million Series B, upsized from a $31 million round, plus $15 million in venture debt. Those two figures are money in the bank. The $140 million is a conditional intention, and the company was careful to describe it that way.

Why the ceiling number ends up in the forecast anyway.

Because the ceiling is the only number anyone gave you. A task order has no value until it is competed. A framework has no value until a site manager orders. So the seller reaches for the one figure the press release contains, and the press release contains the ceiling because the ceiling is what makes the announcement newsworthy. Nobody set out to inflate anything. The number that exists gets used.

Because the vehicle really was hard to win. Getting onto a Navy IDIQ requires a compliant proposal, past performance and, for an early company, a great deal of unpaid work. Signing a global supply framework with Amazon requires surviving a supplier qualification most vendors never finish. The effort is real, so the win feels like a close. Effort is not a forecasting input. Neither is difficulty.

Because the buyer is famous. A logo does the argument for you in a board meeting, and the questions stop. The seats that would normally ask what was actually ordered are the same seats mapped in the hardware buying committee, and a well-known counterparty is the fastest way to get all of them to nod at once.

Because there is no column for it. Most hardware CRMs have stages that run from qualified to closed-won, and a contract vehicle is neither. So it gets filed as an opportunity with the ceiling as the amount and a close date somebody guessed, and from that moment it is indistinguishable from a real deal in every report the company runs.

What this does to a hardware company, specifically.

It hires ahead of demand that never arrives on schedule. A framework or a vehicle produces revenue on the buyer's cadence, not the seller's. A company that staffs delivery against a ceiling ends up carrying capacity through the quarters where the task orders did not come, which is the most expensive mistake available to a business with a physical cost structure. The sequencing question is the subject of hiring the hardware sales function.

It hides the pilot problem instead of solving it. Median pilot-to-production conversion across hardware is 12 percent (IDC 2025, in Hardware GTM Benchmarks 2026). A contract vehicle does not change that rate. It moves the conversion problem one step later and makes it harder to see, because the company now has an impressive agreement to point at while the same handoff sits unresolved underneath it. The anatomy of that handoff is in pilot to production.

It teaches the sales team to stop at access. If the vehicle is celebrated as the win, the next behaviour is more vehicle-chasing, because that is what got applause. The work that produces task orders is unglamorous: finding the individual programme with funded requirements, being on the right side of a specification, and being present when the requirement is written. Nobody announces that work.

It sets a valuation on a number the company cannot control. Ceilings compound in a deck. Two vehicles at $40 million each is a $80 million story that represents zero dollars of obligated spend, and a company that has told that story has to keep telling it. The correction, when it comes, is not a bad quarter. It is a credibility event.

Four rules for forecasting a vehicle honestly.

One: keep ceilings in a separate column that never rolls up. Track them, report them, be proud of them, and never let the number enter the revenue plan. A vehicle is a distribution asset and belongs beside channel access, not beside bookings. The moment a ceiling is allowed into the same total as an order, no reader of the report can tell the two apart again.

Two: forecast the next conditional milestone and its date, not the headline. For an IDIQ that is the first task order solicitation and the date it is expected. For a framework that is the first site, the first purchase order, and who at the buyer has authority to issue it. For a letter of intent it is the specific condition that converts intent into obligation. Each of those has a date and an owner. The ceiling has neither.

Three: name the competitor set on a multiple-award vehicle, out loud. Being one of several selected companies is a public fact and it is usually in the announcement. A forecast that does not state how many other awardees are competing for the same task orders is not a forecast, it is a hope with a number attached. If the answer is unknown, that is the first thing to go find out.

Four: convert access into a design or specification position while the vehicle is warm. The reason a vehicle is valuable is that it puts a seller in the room before requirements are final. In semiconductors that is a design win, and the revenue arrives on the customer's production schedule rather than the seller's, which is the argument in semiconductor GTM. The same logic holds anywhere with a long qualification cycle. Access has a half life. Spend it on being written into a specification, not on a press release.

The version of this that is actually good news.

None of this argues against pursuing vehicles, frameworks and letters of intent. They are among the most durable commercial assets a hardware company can hold, precisely because they are difficult and precisely because they outlast individual champions. Amazon's framework means AutoStore does not have to re-win procurement at every site. Blue Water's IDIQ position means the company is inside the competition rather than outside it. Multibeam's letter of intent, if it converts, funds process development that no venture round of that size would have covered on its own.

The discipline is simply to call each of them what it is. Access, not demand. A shortened path, not a completed one. The companies in this note all described their own agreements accurately in their own releases; AutoStore in particular wrote the disclaimer into the announcement itself. The overstatement rarely comes from the company that signed. It comes from the second telling, in a board deck or an investor update, where the qualifier gets dropped because it makes the slide weaker.

A hardware business that keeps the qualifier ends up with a forecast that is smaller and true, which is the only kind that survives contact with a capital cycle. The rest of the method is on the GTM consulting page, and the earlier-stage version is go-to-market consulting for hardware startups.

Three questions, answered straight.

Is a framework agreement or supply agreement revenue?

No. A framework agreement sets the terms under which a buyer may purchase. It does not commit the buyer to purchase anything. AutoStore's own announcement of its strategic supply agreement with Amazon states that while the agreement defines the terms for further procurement of AutoStore systems by Amazon, it does not contain any purchasing commitments at this time (AutoStore, August 13, 2026). A framework removes procurement friction from a future order. It does not create the order, and it should be forecast as access rather than as pipeline.

What does an IDIQ ceiling actually guarantee a vendor?

Almost nothing on its own. An indefinite-delivery, indefinite-quantity contract sets a maximum total value across all awardees and a period of performance. On a multiple-award IDIQ the vendor has won the right to compete for individual task orders against the other selected companies, not a share of the ceiling. Blue Water Autonomy was selected in August 2026 for a multi-year multiple-award IDIQ with a $40 million ceiling for autonomous open-ocean survey work, and will compete for task orders alongside the other awardees. The forecastable number is the task orders won, not the ceiling.

How should a hardware company forecast a letter of intent or an MOU?

As a qualification event with a date, not as bookings. A letter of intent records that a counterparty intends to proceed subject to conditions that have not yet been met. Multibeam announced a letter of intent with the U.S. Department of Commerce for $140 million in proposed CHIPS research and development funding on July 29, 2026, one of seven letters of intent Commerce announced that month totalling $874 million (NIST). Proposed funding under a letter of intent is not obligated funding. The disciplined treatment is to forecast the next conditional milestone and its date, and to carry the headline number only in a separate ceiling column that never rolls into the revenue plan.

Keep reading.

This note sits inside a published method. Start with the page that ties it together, or go straight to the piece that names your stall point.