There's too much bureaucracy to innovate. A potentially great idea could be snuffed out because it threatens the internal ambition of a middle manager, or the silos required to overcome results in burnout and good devs leaving. Even if a good idea is approved, you need 20 different teams, requiring 20 different meetings, and overall meetings, and then you need status meetings to watch the budget, etc etc.
> Also, strong, small teams usually don't scale well - which is why they sell themselves to megacorps who can
They sell to megacorps to make money, not to scale their product.
I should have been more explicit. Scale is the arbitrage opportunity exploited by megacorps: Buying an upstart at $X means they project that they can make a multiple of $X out of the transaction, based on their ability to scale (or integration with other products). If smaller team could easily make the same multiple of $X, then they wouldn't sell at $X.
Also, strong, small teams usually don't scale well - which is why they sell themselves to megacorps who can (e.g. megacorps are much better when it's time to localize to 20 countries on 3 continents and navigate the legal labyrinth necessitated. All that work requires meetings to coordinate, even if legal ends up only asking the engineer a single question)