The AI computing power story is being spun all the way to 2030
Gartner has laid out the numbers: by 2030, AI data centers will need to earn $8.5 trillion just to cover costs and deliver a 12% return on capital. But actual AI revenue from businesses and consumers will only reach $3.6 trillion—leaving a gap of more than half. Data center investments worth $3.5 trillion are facing impairment charges, and 80% of new cloud providers won’t survive as independent operators through 2030. Meanwhile, CoreWeave is still insisting its $104.2 billion backlog is enough to keep its debt manageable. They’re certainly talking tough.
Put those numbers together, and the upstream cash flows behind the AI computing-power narrative start looking questionable. Stocks riding on expectations for decentralized GPUs and computing-power rentals are priced on the assumption that AI capital spending can keep flowing indefinitely. Once impairment fears take hold, market makers will probably ride the hype to push prices up, then dump their positions—a classic spike-and-fade setup. As liquidity follows the broader market, when AI stocks get derated, money may shift into Bitcoin, making it a safe haven.
For those of you still holding AI computing-power tokens: what’s your cost basis? If you’re lining up on the rooftop, remember to call out your number.
$TAO $RENDER $BTC
Gartner has laid out the numbers: by 2030, AI data centers will need to earn $8.5 trillion just to cover costs and deliver a 12% return on capital. But actual AI revenue from businesses and consumers will only reach $3.6 trillion—leaving a gap of more than half. Data center investments worth $3.5 trillion are facing impairment charges, and 80% of new cloud providers won’t survive as independent operators through 2030. Meanwhile, CoreWeave is still insisting its $104.2 billion backlog is enough to keep its debt manageable. They’re certainly talking tough.
Put those numbers together, and the upstream cash flows behind the AI computing-power narrative start looking questionable. Stocks riding on expectations for decentralized GPUs and computing-power rentals are priced on the assumption that AI capital spending can keep flowing indefinitely. Once impairment fears take hold, market makers will probably ride the hype to push prices up, then dump their positions—a classic spike-and-fade setup. As liquidity follows the broader market, when AI stocks get derated, money may shift into Bitcoin, making it a safe haven.
For those of you still holding AI computing-power tokens: what’s your cost basis? If you’re lining up on the rooftop, remember to call out your number.
$TAO $RENDER $BTC