You seemed to imply that SGI, Sun, etc are gone because they were full stack. I just offered counterexamples of non full stack companies that are also gone. There are simply no evidence that full stack is inherently bad, or vice versa. It's all specific to individual companies.
> Non-full stack companies come and go. That's part of the strength
Or more precisely, that's a benefit of hardware commodity. From the business perspective however, it isn't fun to be a loser, full stack or not.
There is evidence that full stack is hard to keep going. In general computing devices there are only 2 companies left doing it that matter. IBM & Apple.
The non-full stack companies that are gone have been replaced.
The full stack ones have been replaced by combinations of non full stack companies.
New non-full stack companies come in and get big. What new full stack company that matters that has been started in the past decade?
You are contradicting yourself. You say that fill stack companies are hard to keep going and yet you say that non-full stack companies come and go even more easily.
By your logic, non-full stack companies are even harder to keep going.
It's not a contradiction. Both are hard to keep going. Full stack ones are even harder. Hence all but 2 are gone.
The ecosystems for more open systems are also really durable.
Personally, watching for decades the most surprising thing is that MS has kept the desktop rather than losing it to Linux.
Would you say that MS has 'won' the server market because they make more money than any Linux company even though Linux has a largest market share of internet servers?
MS keeping the desktop rather than losing it to Linux contradicts your thesis. You can dismiss it as a 'surprise' or acknowledge that your reasoning doesn't accord with reality.
> Non-full stack companies come and go. That's part of the strength
Or more precisely, that's a benefit of hardware commodity. From the business perspective however, it isn't fun to be a loser, full stack or not.