Arthur Hayes Questions US Insurer Solvency Over AI Debt Exposure

Arthur Hayes has suggested that the US insurance sector is technically insolvent once artificial intelligence-related debt is valued at market prices. This assertion is based on captive reinsurance structures that record valueless commitments as genuine capital.

Forensic accountant Thomas Gober, whose findings were published via analyst Nick Nemeth, identified $1.54 trillion in affiliated reinsurance against $657 billion of surplus. Removing that reinsurance leaves 29 of the top 30 US insurers in a position of technical insolvency.

Three Vermont-based captives reviewed by Gober held only 3.7% of the assets required to satisfy their combined obligations.

What Happens if He’s Right?

As long as the underlying contracts are not tested, nothing changes. According to Hayes in his newsletter, the catalyst would be a series of downgrades on AI data center debt.

Insurers now occupy an increasing portion of the AI data center debt market. That connects their solvency to whether AI laboratories keep buying computing capacity.

Private credit funds have already halted investor redemptions once this year. That indicator runs through the same private credit debt stress that insurers are now facing.

State guaranty funds are intended to backstop failed insurers, but payouts are capped at around $250,000 to $300,000 per policy. Those funds are financed by surviving insurers, many of which depend on the same reinsurance model.

Retirees who hold annuities from these insurers could experience actual losses if the underlying reinsurance is as weak as described.

To Put it Simply…

This is a sequence of dominoes, and the key question is what occurs if the first one tips.

Insurers have employed accounting methods to make their reserves appear stronger than they actually are. If AI companies slow their purchase of computing power, the loans backing those data centers receive downgrades, and insurers holding that debt suddenly require more real capital than they possess.

That is when the accounting trick is revealed. Some private lending funds are already showing comparable stress, an early indicator of what might follow.

If an insurer collapses, the safety net is weak: government-backed funds cover only $250,000 to $300,000 per policy, and that safety net is supported by other insurers who may be equally exposed.

Thus ordinary individuals with retirement annuities could end up absorbing the loss.

This may sound familiar because it mirrors what occurred in the 2008 housing collapse, which the film The Big Short portrayed.

Steve Eisman, one of the real investors featured in The Big Short, appeared on a podcast to discuss this exact research and described it as “a slow brewing scandal which could one day be a great financial crisis.”

What Would be the Outcome for Bitcoin?

Two possibilities remain. Regulators could let insurers fail and face the repercussions, or the government could step in to prevent it.

Hayes has presented this same argument previously. He connected the Federal Reserve’s defense of the yen to additional dollar liquidity. He also compared Treasury Secretary Bessent’s buyback program to Yellen’s 2023 strategy.

The same logic applies here. Whether the government bails out insurers or purchases AI computing capacity directly, both routes expand the money supply. Hayes views that expansion as positive for Bitcoin (BTC).

Whether Hayes is correct depends on disclosures that regulators have kept confidential for years. If the shortfall is even a fraction of what forensic accountants describe, the consequences would extend well beyond crypto markets.