[wasabi991011]

> but somehow companies that make freaking EVs, robots, and rocket ships, and AI can’t possibly be worth that much?

The sky-high valuation of SpaceX is almost entirely related to it's estimated TAM from AI entreprise solutions, not robots and rocketships. HN news comments have a similar bias against the sky-high valuations of OpenAI and Anthropic.

[rayiner]

> The sky-high valuation of SpaceX is almost entirely related to it's estimated TAM from AI entreprise solutions, not robots and rocketships

That’s not how TAM works. The valuation of each business unit isn’t just a simple proportion of its TAM like that. In the SpaceX/xAI merger, which was just a few months ago, the rocket company was valued at $1 trillion and the AI company at $250 billion: https://www.reuters.com/legal/government/how-math-works-175-...

EDIT: To elaborate: TAM is not a valuation for a specific business. It’s a ceiling in the size of the market the business targets. AI has an astronomical TAM because you can sell AI into almost every market. E.g. shipping and logistics is a $10 trillion business. You could sell AI into that market and capture some of that revenue. But if one business has a $10 trillion TAM and another is $5 trillion, that doesn’t mean the valuation of the first business is double the second one.

[jcranmer]

SpaceX's S-1 listed its expected TAM and did break it out by segment. $26.5 trillion to AI, $1.6 trillion to connectivity (Starlink), $0.370 trillion to space launch. https://www.sec.gov/Archives/edgar/data/1181412/000162828026...

[aftbit]

$26.5 trillion dollars. That's more than $3,000/yr from every single man woman and child on the planet. Another way to put it - more than 10x as much as all of the US spends on food every year. For Grok/xAI. That's bonkers.

[tim333]

It's not for Grok/xAI. It's an estimate for the total market for AI at some future time. As one datapoint:

>Morgan Stanley projects a $25 trillion market for AI-powered robots alone by 2050,

It's about 10x current AI spend.