Build or Buy: What to Own in a Hardware Product

Commercial off-the-shelf parts are one of the easiest decisions a team can make and one of the hardest to undo. Buying a ready module gets you moving in a week. It lowers the upfront bill. It feels like risk removed. For the parts that are not your product, it usually is the right call. The trouble starts when a team reaches for off-the-shelf not because a part is a commodity, but because building feels hard, and gradually hands away the part that was meant to set the product apart.

The upfront saving hides a running cost

Off-the-shelf looks cheap on day one because you compare a purchase order to an engineering effort. That comparison is misleading. The purchase order repeats on every unit you ship, and it repeats on the vendor's terms: licensing, per-unit fees, minimum orders, and price rises you do not control. At volume, a bought subsystem can quietly erode the margin the product was meant to earn. The engineering effort is paid once. The purchased part is paid forever.

You inherit someone else's critical path

When a core function lives in a vendor's box, their roadmap becomes your roadmap. You wait for their next release. You work around their integration quirks. You cannot fix the bug that is blocking you, because you cannot see inside the part. The schedule you promised now depends on a supplier who does not share your deadline. Delay stops being something you can engineer your way out of and becomes something you queue for.

The capability you never build

This is the cost nobody puts in the spreadsheet. Every subsystem you buy instead of build is a capability your team never develops. The knowledge stays with the vendor. The next product starts from the same weakness, so you buy again, and the gap widens. A company that keeps buying its hard parts is quietly training itself to be an integrator of other people's technology rather than a creator of its own.

Where your differentiation lives

Your differentiation is supposed to be the thing a competitor cannot easily copy. If that thing is an off-the-shelf module, then it is not yours; it is available to anyone with a purchase order, including your competitors. You own the assembly. You do not own the capability. When the interesting part of your product is a part number, you have no moat.

Buy what makes you the same as everyone else. Build what makes you different.

When a supplier can become a competitor

There is a quieter consideration. The vendor of a critical subsystem sees the system built around their part, and over time they learn the application well. A supplier who understands the full product is well placed to move up the stack toward it. That is not a reason to distrust suppliers, they are essential; it is a reason to be deliberate about how much of what makes your product distinctive sits inside a part you do not control.

Own the core, buy the rest

None of this is an argument to build everything. Building the commodity parts, the ones you could buy from anyone at the same quality, is just ego and wasted time. The discipline is knowing which is which. Buy what is undifferentiated. Build what is the soul of the product, the part that defines why a customer chooses you and not the alternative. Get that division wrong in the direction of buying, and the short-term speed can quietly cost you the long-term position.

So before the next build-versus-buy decision, it helps to ask a sharper question than "what is cheaper this quarter?". Ask "is this the thing that makes us who we are?". If it is, it is usually worth owning, even when owning it is hard. The teams that do best over the long run tend to be the ones that decided, early and deliberately, what they would not buy.