On the other hand I have no trouble coming up with examples of hardware companies that did well: Apple, Nvidia, Intel... but those time scales are titanic.
Not sure they are great examples. Intel is going through a lot of pain now (but it's been very successful in the past). Apple was in a terrible situation before its reverse acquisition by NeXT (NeXT paid one Steve Jobs for all of Apple and got $400 million in change). Nvidia got insanely lucky with Bitcoin, then with AI. Its original plan was to make 3D accelerators for gamers and, maybe, engineering workstations.
All of them were a couple wrong decisions away from doom multiple times.
Think Commodore, who made one of the most popular computers ever, only to be mismanaged into the ground.
Nvidia got lucky by building a product that just happened to be amazingly well suited to a technology that would emerge 20 years after they were founded. Credit where it's due, they developed CUDA and gave universities gobs of cash/hardware to train students to leverage CUDA for machine learning and later, AI. But if not for AI, Nvidia would still be a video card designer with a market cap of maybe 5% of its current valuation.
It's difficult to call them a hardware company in the sense that Intel is. They only do designs of hardware, and a lot of their value comes from the software they designed to leverage their hardware in ways beyond their initial purpose.
The main issue now is competition for capital: the majority of tech investors regard software startups as much safer paths to huge returns. Whether they’re right or wrong in any individual case is a separate issue; they’re playing the percentages.