Credit is drying up: AI has soaked up global savings and Elon Musk can't find funding

Musk cannot secure funding for $40 billion in GPUs while global savings run dry in AI. With bonds at 5%, credit is closing for everyone.

English · Original discussion in Spanish · Published

Credit is drying up: AI has soaked up global savings and Elon Musk can't find funding
The credit tap is closing: AI has already drunk the world's savings

The news is a $40 billion debt to buy GPUs. The trigger: Elon Musk cannot find funding for his company. And behind both, a thesis that runs through the whole conversation: cheap capital is over. With US bonds at 5% and rates rising in half the world, available savings have already been poured into artificial intelligence, and what's left is not enough for mortgages, new funding rounds, or the next big score.

A phrase that appears early on sums it up: "Credit is finished because we've already consumed it all." It's not a quip. SpaceX is valued at $2 trillion in market cap, and last year, just by renting out its spare processing capacity, it made $6 million in profit. With those figures, the market shutting the tap on Musk says a lot about the state of the market. If the richest man on the planet can't borrow, neither can the rest.

The Cantillon effect of the $40 billion

The operation has a technical name and a popular one. The technical: money issuance that benefits those closest to the hose and hurts those who arrive late. The popular: the Cantillon effect. The $40 billion won't produce new chips or fund factories; they will buy out the entire warehouse, driving up chip production costs worldwide and leaving the rest of the planet with the bill.

Some point out that the move recalls what Google did more than a decade ago with its class B shares, and that Meta is now preparing a class C share issuance to raise funds without diluting voting rights. The difference, they note, is the scale: never before has so much money been mobilized for an asset with no distributed profits to offset it.

Why doesn't AI return the money it receives?

Because the product exists, but it doesn't make money. That's the central contradiction. The empirical results are there — some argue we're on the verge of a paradigm shift comparable to electricity or oil — but the companies would need trillions of dollars in profits to repay what they've borrowed. Meanwhile, the global savings that could have gone to other things have been locked up in GPUs that will be obsolete or burned out in three years.

The most pessimistic calculation doesn't speak of a correction, it speaks of investment fund bankruptcies. And a suspicion hangs in the air: trapped investors are pushing to stretch the rubber band until the IPO and leave the corpse to someone else. The problem is arithmetic: there isn't enough money in the economy for two companies to each raise more than two trillion dollars.

Will credit revive if rates fall?

The path being pointed to is geopolitical: if Iran is sorted out and the war in Ukraine ends, inflation is contained, rates fall, and credit returns. That's the optimistic hypothesis. The pessimistic one is that money is already being printed to prop up the system, with inflation soaring without a corresponding rise in wages: robbing the average worker so credit keeps flowing.

And a scene that closes the circle: Nvidia lends money to buy GPUs from Nvidia. The buyer pays with the seller's money, who profits from selling what he himself financed. When that breaks, they say, the bang will be heard in Vladivostok.

Also available in: Deutsch

Summary of a discussion on Burbuja.info - Foro de economía, actualidad y política., translated from Spanish and reviewed before publication. Read the full discussion (21 replies).

More summaries

All summaries in English →

-
Back