Even setting that aside, not everything is or should be a land grab. It's notable that all the examples you provided — Amazon (at least, its initial online store product), Uber, Facebook — are all B2C plays and I don't think that's a coincidence.
I'd argue most b2b/enterprise software is a new version of something that already exists or addressing a need that already has a market. Business model is also very clear, there is very little network effects usually other than reputation and customer proof. Yet most the startups not even close being profitable.
In my mind most software products are differentiated so in the end the main success comes from getting the differentiation right for the market, not outspending the competition.
That is not today's environment.
When it's the focus, profitability can be easily achieved somewhere between day 1 and month 18.
The trick is when you're trying to take risks and innovate. It took Amazon a long time to be profitable. It took Uber a long time to be profitable. It took Facebook a long time to be profitable.
When it's a land grab - when you're racing against other companies in a new market like AI - you need to burn money fast to run fast. Can't take a year in private beta.