OpenAI, Anthropic, SpaceX to trash stock market IPOs for 2026

Venture capital works by turning startup investments into dollars — eventually. It’s not their own money — it’s their investors’ money. And those investors are getting impatient for the payoff. One…

Pivot to AI
I get stock market valuations are always ridiculous, but what in the fuck does it *mean* for OpenAI to be "worth" $800 *billion*. Like, what the fuck did Altman do, go individually pickpocket $100 from every person on the goddamned planet?

@xgranade Are you asking literally? I'm going to give an explanation, but sorry if this was just rhetorical and you're not actually looking for an explanation.

Valuation just means the expected value that the company will produce over all time, based on what the most recent round of investors paid.

With any company, there's some non-zero probability that it will go bankrupt, and some non-zero probability that it will make varying amounts of money, and people who invest in the company do so based on their expectation of what that value will be.

So no, it doesn't mean he pickpocketed $100 from every person on earth. But it means that if you average out all of the possible expected outcomes, the most recent investors expect him to make $100 from every person on earth, over all time; or an equivalent amount of money from a smaller number of people.

Given that the rate of revenue they are making is $25 billion a year, this number doesn't seem so far fetched; now, they're also spending money far faster than that. But a lot of that money they are spending is to build the infrastructure used to meet expected future demand, so the theory is at some point they will be making more than they are spending.

Is that true? Well, that's hard to say; there is definitely a lot of signs of a bubble that will burst. But also, they are providing a service that a lot of people really are willing to spend a lot of money on, so it's not all smoke and mirrors; it's just a question of whether those lines will cross before they run out of money and people willing to invest more in them, and what the actual profitability ends up being after they do cross.

I think it's important for us AI skeptics/haters to realize: there are people who do find value in these services. We might hate it, we might be skeptical of whether the value people find in it matches up with what they're willing to pay in the long term, as some of the recent value and spending may be due to hype and FOMO. But yeah, people are really spending money on this, there is a non-zero chance that they will become profitable and make a significant amount of money, and while there may be some irrational exuberance and a bubble that bursts, this is going to be a big business long term.

@unlambda @xgranade At current prices they’re losing money on every inference just from operating expenses, not cap-ex, and it’s speculated that the break-even point is anywhere between 10 and 100x what they’re charging, which is already too high for the vast majority of the public, who, according to surveys, largely hate it. This won’t be “big business,” unless a lot of things change at once in ways that don’t seem likely to me.