Your Prototype Works. Now Make It a Product.

A working prototype is a real milestone. It is also where many hardware teams start making the wrong assumption, that the hard part is mostly over.

The prototype proves the design works. It does not prove you can build it a thousand times, with real suppliers, real lead times, real quality constraints, and real margins. That gap between one working unit and thousands of shipped units is where many hardware startups break down.

One of the biggest reasons hides in the BOM.

A finished BOM?

When the design freezes, most teams freeze the BOM with it. The part numbers stop changing, so the list feels done.

But every line on that list is a bet on a moving market. That the supplier keeps making the part. That stock is there when you need it. That the price still fits your margins. That the lead time still fits your schedule. That a second source exists, and that switching to it will not eat your launch date.

During prototyping, those bets almost always pay off. You need one of each part, and you need it this week. Volume is what puts them to the test.

That is why so many scaling problems get filed under supply chain when they really started as design decisions, months earlier. By the time procurement finds the issue, the design is already expensive to change.

One part stopped the build. Almost.

One of our customers, a mid-production hardware team, had a Micron 128Mbit SPI NOR flash on their BOM: MT25QL128ABA1ESE. It looked safe. It cleared every check most teams would have run at the time.

Through the second half of 2025 it traded around $1.60, with roughly 21,000 units in distributor stock. Then the market moved.

Since May 2026, there has been zero stock anywhere. While scattered broker offers remain, the cheapest today is $16.05. Ten times the 2025 price in quantities you cannot verify.

Any build carrying that line is stopped. Not delayed. Stopped.

Except this one wasn't. Our alternatives agent scanned the market and returned parts that matched the original on fit, form, and function. All active. All in stock that day.

The Macronix drop-in starts at $1.93. Roughly what the original cost in 2025, before any of this happened. The team swapped it in and kept building. Same design intent. No schedule slip. That is the whole difference between a sourcing surprise and a product decision.

None of this happens because teams are careless. It happens because a frozen list cannot tell you when the market changes.

A BOM is a living thing

Prices climb. Lead times stretch. Parts go end of life. Second sources appear and disappear. The part numbers hold still while everything underneath them moves.

That is what our BOM optimizer was built for. Every line is reviewed right after design for the risks that decide whether you can ship: cost, availability, lifecycle, second source coverage, compliance and tariff exposure, and how hard a swap would really be.
We watch the market 24/7 and keep ranked, buildable alternatives ready. So when something moves, you make a product decision instead of taking a production delay.

Validation happens once. DFM happens once. These are checkpoints in the product development lifecycle. A BOM that cleared every check in March can be exposed by August without anyone touching the design.

If a risky part gives your product a real edge, keep it. Just know early whether replacing it is a one-line ECO or a three-month redesign.
And that knowledge should live with you, not only with your brokers, distributors, or CM. When your team owns it, sourcing tradeoffs become part of product decisions instead of surprises from outside.

The prototype proves the product can work. The BOM proves whether you can ship it.

The next generation of hardware companies will not be the ones that avoid every risky part. They will be the ones whose BOM is alive: watched, ranked, and ready to move the moment the market does.