Why Hardware Is Back in Vogue: Building Defensible Moats in the AI Era

Hardware outcomes seem binary.

Hardware investing has produced some of our biggest wins at Founder Collective – Shield AI, WHOOP, Verkada – and at the same time some of our larger losses. It’s about 10% of our overall portfolio.

Before I was a VC, I built a hardware startup that was eventually acquired by 3M. I lived the very real binary risk of the physical world. Not to mention getting the BOM (build of materials) to an acceptable cost level, constantly improving usability and form factor, and a myriad of other seemingly endless issues.

Right now hardware is back in vogue. Investors are realizing that surviving real-world friction creates a defensible asset that an AI coding agent can’t replicate.

As I start to see more hardware start-ups I’m both intrigued and cautious. The hard part is actually founder psychology.

To survive the hardware “valley of death,” you need a very unusual founder profile. A hybrid.

If a founder is purely sales-driven, they often lose their mind. Hardware requires a long, multi-year R&D slog before you have a single unit to sell. Highly commercial founders often lack the patience to survive that desert.

If a founder is purely an engineer, they might build a physically perfect, flawlessly spec’d machine but constantly tinker and never get it to market.

If you are building in atoms, the moat can be massive. But it requires a founding team that balances deep technical patience with aggressive commercial execution.

It’s not for the faint of heart. But I continue to invest as I think in some ways it’s what VC should be investing in.