The projections would go something like $0 two years ago, $5k last year, but then $100k next year, $5mm year after that. One day? $1B, easy.
There'd be precious little evidence to support the exponential growth hypothesis, while having 10 assigned engineers earning as much as their peers in more predictable domains.
It's easier to justify such ideas in a big, rich company when the risk free rate of return is near-zero and the profit center of the company is compounding (the growth in Ads justifies the negative ROIC everywhere else).
But when your profit center is wavering and showing worrying signs of stalling out (as Google and Meta did in 2021/22), the calculus shifts dramatically.
When you manage a company not for "growth" but for "value" (which is an inevitable part of the corporate cycle), these projects and the employees behind them are decreasingly seen as a potential source of future profits and more as a waste of increasingly valuable cash.
The future earnings should go up with inflation, but not the discount rate. The government adjusts their interest rate to maintain the rate of inflation around 2% or so. The government interest rate is often around 5%, but can vary from 0 to 20%. The discount rate (or technically the cost of capital) is the gov't interest rate plus 4% or so (the higher the better for investors).