Key Notes
- Hayes argues an AI investment bust could trigger liquidity support that benefits Bitcoin.
- Excess computing capacity could lower the cost of running AI agents.
- His FLOP network remains in development, with a 2027 mainnet target.
Arthur Hayes says an AI investment bust could eventually benefit Bitcoin, even as it leaves data-center investors facing losses. The BitMEX co-founder’s argument connects two outcomes: government support could restore financial liquidity, while excess computing capacity could make AI agents cheaper to run.
Hayes outlined the computing side of that thesis in his October 7 TOKEN2049 speech in Singapore, according to a published transcript. He also promoted FLOP, his developing network for AI-agent payments and inference. The commercial connection matters: Hayes is describing a future in which his own project could benefit.
His skepticism concerns the economics behind the infrastructure boom, rather than whether AI has useful applications. He argues that subscriptions to OpenAI, Anthropic and xAI do not reflect the full cost of providing their services, leaving investment dependent on expectations of much greater future demand. Those are Hayes’s assessments, rather than an independently verified comparison of the companies’ current finances.
How an AI Bust Could Benefit Bitcoin
Hayes develops the financial argument in his Safety First essay. His concern is that debt remains payable even if the demand supporting data-center leases weakens. Losses could then spread beyond AI developers to the institutions financing the buildout.
He sketches two potential responses: the government becomes a major buyer of computing capacity, or it supports financial institutions exposed to troubled AI debt. Hayes expects the resulting monetary expansion to favor Bitcoin. The bullish part of his scenario therefore comes from the rescue and its financing, rather than from the initial collapse itself.
A bailout is not an announced policy, and neither its size nor its effect on Bitcoin is assured. A disorderly sell-off could initially force investors to raise cash by selling liquid assets, including cryptocurrency. Even if support follows, higher liquidity would compete with other influences on prices, such as interest rates, regulation and investors’ willingness to take risk.
That distinction builds on the earlier thesis we examined: a successful technology can coexist with poor returns for the companies and lenders financing it. Growing AI usage alone does not establish that every facility will earn enough to meet its obligations.
Cheaper Compute Is the Other Side of the Argument
At TOKEN2049, Hayes argued that excess capacity could drive computing costs lower and support more capable, widely available agents. An infrastructure correction would not erase the software built on top of it. Instead, surviving users could gain access to resources that were expensive during the investment rush.
The practical question is how much of any cost reduction reaches customers. An agent service still needs software, storage, networking and support, alongside inference. Cheaper GPU time could improve its economics without making the full service nearly free. Utilization and reliable delivery also matter to a business trying to run persistent agents.
Where FLOP Fits
In The Book of Genesis, Hayes presents FLOP as a currency connected to the resource agents need to operate: compute. His proposed market would let agents request inference from independent providers, with suppliers receiving payment for useful work.
The project’s draft whitepaper describes a dedicated blockchain linking agents, compute providers and validators. Its native FLOP token would pay for inference and support settlement. The document lists a fourth-quarter 2026 testnet and a first-quarter 2027 mainnet as targets, while noting that the definitive specification is unfinished.
Those plans should be distinguished from a proven commercial network. Verification, dependable supply and sustained customer demand still need to work together. A token can coordinate incentives, but it does not by itself demonstrate that customers will prefer the service or that providers can deliver inference competitively.
Agents Have Competing Payment Options
FLOP’s proposition also enters a broader debate about how agents should transact. As our BlackRock coverage explored, stablecoins offer another possible route for machine payments. A dedicated compute token would have to justify the additional exchange and price-management steps it introduces.
Hayes’s two bets can consequently succeed or fail separately. A liquidity response could lift Bitcoin without establishing FLOP as an agent currency. Equally, a useful compute marketplace could find customers without the predicted financial crash. The test for FLOP will be the service it delivers, while the bailout thesis remains a conditional view of how an AI credit downturn might unfold.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
