There is an upwelling of interest in physical technologies ("deep tech") since they might have more impact on the real world. But many talented investors see the space as a money pit that can't produce software-like returns. I think there are similar issues with trying to buy complex technologies.
Cost Plus vs. Fixed Price
There are infinite variations in contract structures, but most buying happens through cost-plus or fixed-price contracts.
Cost plus is where you tell someone to build something for you, and then they charge you the cost plus some profit margin. You might also know it as "time and materials." The buyer takes the risk.
Fixed-priced contracts are when the parties agree to a final price beforehand, putting the risk on suppliers.
You might wonder why any buyer would ever agree to cost plus. There are a few reasons:
Lack of information about the cost. Vendors will be reluctant to provide a fixed price. They might inflate it to a worst-case scenario if you twist their arm.
Monopsony. A supplier can fail because the buyer is most of the supplier's business, and they misjudged the cost of fulfilling a contract. Companies like Boeing might get a fixed price bid initially but then have to buy the supplier out to maintain their parts supply.
The vendor has no competition.
Cost-plus can work under certain conditions. Oil companies purchase a significant portion of services (like drilling rigs) on a cost-plus basis. It is successful because contract lengths are short, and there are many buyers and suppliers. You can ditch a contractor after drilling one well if they are awful. The multitude of buyers supports many options. Suppliers want to earn the next well and try their best. Fixed price usually doesn't work because there is massive variation in cost between wells, and drilling requires coordination between as many as fifty contractors. Bad luck or a mistake by one vendor punishes everyone even if they perform well. Some drilling is more predictable and can support some variation of fixed cost. The drilling rig company or a supervision firm will act as the prime contractor and coordinate vendors in these cases.
Examples From the US Government
Governments have many famous failures when buying big-ticket items like ships or rockets. Cost-plus examples like NASA's Space Launch System are the most well-known, but there are fixed-price tragedies, too. The KC-46 tanker program is years behind schedule. Theoretically, Boeing is on the hook for the cost overruns, but the Air Force is missing out on a critical capability as it struggles with ancient, unreliable KC-135 aircraft. It has reached the point where Boeing is holding the Air Force hostage and won't finish the aircraft without more money to fix some of its issues.
These projects are technically challenging, which makes estimating their cost beforehand difficult. The lowest bidders often suffer a "winners curse" where they misjudge and underbid. There are successful examples, like NASA's commercial resupply and crew programs.
Making Fixed-Cost Work
Some strategies can overcome contracting challenges. First, the buyer must have talented professionals. NASA makes bidders submit extremely detailed packages, and its engineers evaluate if the technical approach in each proposal is sound. The contracts are also broken up into shorter stages to reduce risk. NASA will award contracts to many companies for preliminary studies, narrows the field to provide detailed engineering, and finally awards a production contract. It has also made a point of spending extra upfront to have multiple vendors for critical systems. One vendor can't hold NASA ransom for more money if they fall flat on their face.
There is no avoiding that successful contracting requires a lot of skill and effort from the buyer.
Summing it Up
Buying and developing complex technologies is always fraught with risk. Situations with knowledge asymmetry or incentive misalignment between parties are recurrent. Both sides must have adequate technical knowledge. The buyer must promote competition through the process and into production or risk becoming a hostage. I think some of these lessons also apply to deep tech investing. More on that later!