Arthur Hayes has stated that he’s popping out of retirement to guide Flop Labs, a brand new challenge centered on a token known as $FLOP that goals to work as a sort of forex for AI brokers.
The announcement doubled as a protection of the transfer: Hayes, who has spent months warning that AI funding is a bubble, argued the surplus sits within the debt piling as much as construct knowledge facilities, not within the underlying expertise his new enterprise is constructed on.
A Honest Launch and a Compute-Backed Token
Hayes described $FLOP as “meals in your AI agent” and stated the token would launch and not using a presale or enterprise capital funding. “100% honest launch,” he wrote, including that he expects a “huge airdrop in This autumn” adopted by a genesis block within the first quarter of 2027.
Flop Labs’ announcement describes the community as a proof-of-useful-inference protocol. Its said aim is to offer AI brokers a local forex for getting computing energy and storing recollections. The challenge makes use of floating-point operations, or FLOPs, as the idea for its financial mannequin. Miners would supply computing energy and obtain $FLOP by means of block rewards and inference funds.
Validators would confirm that miners delivered the computing work requested by brokers. They’d additionally retailer agent recollections and obtain $FLOP by means of block rewards and inference funds.
On their half, AI brokers would spend the token on computing and reminiscence companies, whereas neighborhood companions may obtain $FLOP primarily based on community exercise.
Flop Labs stresses that the community has not launched and stays underneath improvement. Its present design can change, and the challenge makes no ensures about receiving tokens or earning money.
The AI Bubble Argument
Earlier than saying Flop Labs, Hayes had in contrast the AI buildout to the 2008 housing disaster, arguing that lenders, personal credit score funds, and governments are financing knowledge facilities on the belief that demand retains climbing with out restrict.
He expects AI capital spending to gradual in 2027 earlier than contracting, which he stated would pressure governments into bailouts bigger than people who adopted the 2008 monetary disaster, sending new cash into crypto markets and doubtlessly pushing Bitcoin towards $1 million.
When requested why he would construct an AI challenge whereas calling AI a bubble, Hayes stated the surplus lies in debt used to fund knowledge facilities and within the shares of hyperscalers and frontier labs that aren’t but worthwhile, not in agentic expertise itself.
“Worth is what you pay, worth is what you get,” he wrote, including that the compute overcapacity constructed on borrowed cash strengthens his case for Flop Labs.
Hayes’s guess on an agent-native forex additionally arrives towards skinny real-world utilization elsewhere. Analyst Jamie Coutts reported on August 12 that day by day settlement quantity on x402, the Coinbase-built fee protocol for AI brokers, is down 93% year-to-date, with the seven-day common round $41,800, nicely under the $800,000 to $1 million peaks seen in late 2025.
Coutts known as the drop a “actuality examine” on claims that the agentic financial system has arrived, although he expects quantity to climb once more within the fourth quarter.
The publish Arthur Hayes Known as AI a Bubble, Now He’s Launching an AI Challenge appeared first on CryptoPotato.