Meta's CEO just admitted the company's $200 billion AI bet was made on faith, not evidence.
Mark Zuckerberg says Meta committed as much as $200 billion to AI infrastructure before anyone could tell him it would actually work. His own chief AI officer went further, saying it was not clear at all whether the spending would pay off. The admission surfaces as Meta's free cash flow fell 91% in the second quarter. That is about as plain as anyone at Meta has been about how uncertain this bet really is.
Companies usually dress up massive capital spending as foresight. Here, Meta's own leadership is describing it as a wager made in the dark, after the money was already committed. For a company whose last unproven bet - the metaverse - burned tens of billions with little to show, that framing should concern shareholders more than it reassures them.
Meta is not the only tech giant spending first and asking questions later on AI, but it is unusual to hear the people writing the checks say so this plainly.