Arthur Hayes put a number on the next cycle: $125,000 by the end of 2026, $1 million by 2030. The Maelstrom CIO's Oct 8 thesis says every mega bull run begins by scaling a wall of worry. In 2017 it was the block size wars. In 2020, Covid. In 2023, FTX plus rising rates. For 2026, he nominated the FUD around AI breaking the encryption that secures crypto, the same anxiety Justin Drake voiced this week when he urged the industry into "bunker mode" on ECDSA. Hayes ended it with his usual flourish: money will be printed, bags will pump.
The mechanism underneath the number is the interesting part. Hayes frames the AI buildout as a credit event in waiting. AI borrowing in US leveraged finance hit roughly $88 billion in 2026 against $20 billion in early 2025, according to Reuters, and Apollo estimates the AI push needs more than $2 trillion in additional investment-grade debt. The mismatch is the clock: AI hardware turns obsolete in two to three years while the loans funding it run five to six. When data-center revenues stop covering the buildout, banks and private-credit funds eat the losses, and central banks answer the only way they know how. In that world, a fixed-supply asset becomes the escape valve. The bet is not on adoption. It is on the printing press.
Here is the tension worth arguing over. The September FOMC minutes said most officials still see another 2026 hike as appropriate, and the 10-year sits near its highest since 2002. Hayes is betting the Fed blinks when credit cracks. One of them is misreading 2026.
So: is this cycle's real wall of worry the AI encryption scare, or is it the Fed refusing to blink? $BTC $ETH
The mechanism underneath the number is the interesting part. Hayes frames the AI buildout as a credit event in waiting. AI borrowing in US leveraged finance hit roughly $88 billion in 2026 against $20 billion in early 2025, according to Reuters, and Apollo estimates the AI push needs more than $2 trillion in additional investment-grade debt. The mismatch is the clock: AI hardware turns obsolete in two to three years while the loans funding it run five to six. When data-center revenues stop covering the buildout, banks and private-credit funds eat the losses, and central banks answer the only way they know how. In that world, a fixed-supply asset becomes the escape valve. The bet is not on adoption. It is on the printing press.
Here is the tension worth arguing over. The September FOMC minutes said most officials still see another 2026 hike as appropriate, and the 10-year sits near its highest since 2002. Hayes is betting the Fed blinks when credit cracks. One of them is misreading 2026.
So: is this cycle's real wall of worry the AI encryption scare, or is it the Fed refusing to blink? $BTC $ETH