The obvious solution is to promote them. If you don't reward loyalty don't expect loyalty.

Indeed, with a big BUT: you don't want to overvalue loyalty in respect of competence. You don't want to have a boss that is not good but "is here since forever".

Instead, reward them economically. Everytime the company takes a leap forward, make clear to them that they were important for the process, and share some profit.

And make clear that being important in a moment doesn't automatically mean they will be important in the future: they will have to compete on results, like everyone else.

Not everyone is great at the level above and even if they are, sometimes it is not always a good fit for them. I've had more a good few engineers ask to go backwards when they've been promoted - each for their own reasons but nearly always around the theme of being in a position of leadership / seniority.

That's peter principle 101

But if you reward loyalty over skill then don’t expect skill to stick around.

Some people have 5 years of experience and some people have 1 year of experience 5 times.

Since we are a bit nitpicking here, I'd also say don't reward skill, but actual value instead. You can have someone very skilled who is all smarty-pants and derails anyone wanting to actually, you know, do something.

Thus I'd give the previous poster some leeway too, and read their (but also your message) as: there is an implied basis of skill and value you expect early employees to continue bringing to the table, but when that is satisfied, reward their loyalty too (either with promotions or money or status or stock or...).

There’s a common scaling heuristic, related to Greiner’s growth model, that organizations need to fundamentally change how they operate as they grow. I recall numbers that every time your organization triples you need to change how you do things.

Part of this is communication overheads, and as op points out, the need, and ability, to specialise in a larger organisation.

That’s something I’ve never thought about. It’s a valid point and somewhat understandable.

A hard change is the required skills mix.

Early stage you want the generalists who can do everything, move fast and probably break things.

There's an inflection point where you want to stop breaking things and for that you need specialists. Experts in scaling, security, optimisation and code purists.

Finding new roles for your generalists at this point could be hard, even harder will be having to let them go. It's possibly something you should consider at the start and give them the ability to vest and leave for a new greenfield. Alternatively find them a role as an architect/lead where it's their responsibility to be across everything and able to bridge between teams because they have your domain and institutional knowledge.

This can go both ways. Seniors who in earlier days could focus on "the real work" may be expected to increasingly spend their time and attention on other concerns. To CTO or not to CTO, ain't that always the question...

You are not wrong, but I would like to throw in some caveats from my experience:

"Expert" management was brought in who didn't know the industry and didn't know the company and didn't respect the people who were there, who had been eating their own dog food for years. I saw so many stupid decisions being made; I was sidelined and ignored. And then, the corporate bullshit started creeping in where management were lying to their own employees. So, yes, if my opinion stops mattering, it is an unjustified demotion, especially when the layers of management brought in over me clearly don't know WTF they are doing.