All that matters is growth — if you have product-market fit, you'll get users. Initially, investors are looking for a usage curve that goes steeply up and to the right. Eventually, you'll figure out how to earn revenue from those users, either by charging them or through a third-party who wants access to those users (like an advertiser or another company who will pay referral), so the revenue curve will follow the user curve.
And at some point after that, you'll reach a magic point where your costs are spread across enough users that you'll start to turn a profit.
This has led to all kinds of weird accounting. Privately held startups almost never talk about real, GAAP profitability. They talk about positive unit economics. (That's nice — they're not losing money on each sale!) They talk about being cash-flow positive. (That's better — they're taking in more money from operations than they're spending!) They talk about being profitable if you ignore that pesky stock-based compensation. (Which means they really might be on to something!)
This is how companies like Twitter, Box, and Square are able to go public at valuations worth billions without ever having turned an annual profit in their history. All of those companies are over five years old. Twitter is ten years old.
But one Silicon Valley VC, Chamath Palihapitiya, thinks this conventional wisdom is all wrong.