According to insights shared by Woofun AI, Arthur Hayes argues that Anthropic, OpenAI, and SpaceX are citing "safety-first" principles to slow down AGI development, but this is actually a smokescreen hiding their harsh economic reality of failing to turn a profit. The narrative of building "Silicon Gods" is crumbling, as the market's true demand lies in Chinese-priced models costing just one percent of American alternatives, not expensive US technology. This supply-demand mismatch is shrinking computing power demand, which in turn triggers a massive debt crisis lurking beneath the surface.
Digging into the financial truth of AI giants, it becomes clear that profitability claims are under siege from Chinese pricing. As the third quarter draws to a close, Anthropic's failure to go public has sparked intense market skepticism about its financial disclosures. Hayes is eager to examine its forthcoming S-1 filing to determine the actual cost of delivering every single token to users, and whether the base of paying customers is expanding or contracting. However, all questions about cost structures and profit margins are being deflected with the convenient excuse of "safety first."
In reality, market appetite for AI is robust, but consumers gravitate toward Chinese models that cost just one percent of their American counterparts. When US industry players accuse Chinese products of poor quality or dependence on distillation models, the market doesn't buy it—it simply pursues the cheapest intelligence available. As a result, AI giants have pivoted to claiming they're pausing research out of concern for human safety, when in truth they're angling for government regulation and funding to preserve their inflated pricing systems.
The computing power debt crisis is spreading rapidly, weaving a complex web of trillions in liabilities and off-balance-sheet guarantees. The massive computing needs of top AI labs like Anthropic and OpenAI support over $1 trillion in investment-grade debt, plus hundreds of billions in lower-credit-rated loans. Since these labs collectively generate zero profit, they must rely on profitable tech firms such as Nvidia (NVDA.US), Broadcom (AVGO.US), Alphabet (GOOGL.US), and Microsoft (MSFT.US) to back their debts with off-balance-sheet guarantees for data center leases and chip purchases.
Future procurement of chips and hardware hinges entirely on AI labs continuously training cutting-edge models and processing inference requests for clients. If "safety first" becomes the guiding principle, spending on training new models will retreat from its peaks as companies shift focus to improving the efficiency of converting electricity into intelligence, which directly reduces customer spending on computing. In essence, "safety first" is a mechanism to destroy computing demand, thereby shaking the very foundation of the entire debt chain.
The substance of safety-first is demand destruction, which triggers debt default risks. If AI capital expenditures were financed through operating cash flow, the risk would be manageable, but the reality is that trillions in debt remain outstanding. Once AI labs stop consuming computing power at expected scales, the prices of such debts will plummet. The key concern is that speculative buyers of these debts are heavily leveraged, holding large volumes of junk obligations. Woofun AI compiled data shows this leveraged debt structure leaves the market extremely fragile, where any contraction in demand could set off a chain reaction.
The real core question is who ultimately bought these debts and whether they used leverage in doing so. The answer is clearly yes—speculators piled in with high leverage on these low-quality assets propped up by the AI narrative, planting the seeds of a future crisis. The ultimate bagholders are millions of American insurance policyholders who have indirectly bet on the AI story, only to face the risk of insolvency. Nick Nameth on Substack has thoroughly exposed this scam: if AI-related debts were marked to market fair value, a large portion of the US insurance industry would already be technically insolvent.
In a global economy dominated by fractional-reserve banking, this raises the core investment question. Insurance companies sell life insurance and annuity policies, investing premium funds into AI-related debt to chase high yields. However, once AI debt depreciates, insurers' balance sheets suffer severe damage, leaving policyholders exposed to massive losses. This risk-transfer mechanism makes ordinary citizens the ultimate victims of financial elites' speculative gambles.
The US government now faces a binary choice: either act as the final purchaser of computing power in the name of national security, or print money to bail out loss-ridden insurers. Whichever path it takes, bitcoin holders and crypto investors emerge as winners. If the government ignores market signals and pours funds into developing commercially unviable "Silicon Gods," it will need to print money to finance such non-productive spending, which inevitably spawns more financial speculation and pushes bitcoin higher.
If the government opts to rescue the insurance industry, it will print money to absorb bad AI debt, expanding the money supply and likewise driving bitcoin prices up. This dilemma makes monetary expansion an unavoidable outcome, providing a solid macroeconomic foundation for crypto assets.
The political narrative revolves around confronting China and elite rent-seeking. Invoking China, the US government can justify almost any action, much like the global war on terror launched after 9/11. This time, the manufactured adversary is China offering affordable AI products. AI leaders have successfully convinced Trump and his advisors to ignore the market-proven reality that AI businesses don't make money, as well as voter opposition to new data centers. To beat China, the US must implement state socialism within the capitalist system, pouring even more funds into building "Silicon Gods."
This narrative argues that America possesses the most inclusive and fair culture globally, and AGI must never fall into the hands of a non-Judeo-Christian civilization. Therefore, trillions of taxpayer dollars are handed to Elon, Sam, and Dario for research at xAI, OpenAI, and Anthropic. This hubris evokes Icarus flying too close to the sun, doomed to crash and burn.
Macro data and monetary policy reveal a mix of rate hikes and bank balance-sheet expansion, with liquidity conditions remaining accommodative. As of June 2026, US nominal year-over-year GDP growth stands at 6.6%, while the effective federal funds rate sits around 3.6%. Treasury Secretary Bessent keeps issuing more short-term bills; the government earns a 3% spread on this debt issuance, but savers bear the losses. If the fiscal deficit stays under 3%, the debt-to-GDP ratio declines.
However, since July 2023, US monetary policy saw its first rate hike, with the Fed unanimously raising the policy rate by 0.25% at last week's meeting. The Fed is no longer growing the money supply, and the RMP short-term Treasury purchase program has been halted since August 14. If the government pursues a computing power procurement plan without the Fed lowering funding costs or expanding its balance sheet, massive debt issuance would push rates higher, stoking voter anger. Thus, Trump and Bessent need at least seven FOMC members on board to ensure the plan's viability.
Meanwhile, commercial banks have created hundreds of billions in new money by expanding total assets, backed by relaxed liquidity regulation constraints. After this 0.25% rate hike, banks' excess reserves parked at the Fed now earn an extra $7.5 billion annually in interest, fueling new lending and financial market speculation. Combined, the net effect remains stimulative, sufficient to support fresh debt issuance for AI computing infrastructure.
The captive insurance scam reveals how private equity became the buyer and how fictitious capital buffers operate. After the 2008 global financial crisis, the private sector deleveraged and the Fed cut rates to near zero. The classic play for private equity was leveraged buyouts of mature companies, cashing out dividends, then relisting them. As the cost of capital rose, PE leaders sought long-term funding pools, and insurance companies came into focus.
In 2025, private equity and venture capital assets under management surpassed $15 trillion. PE titans acquired insurance companies, appointed themselves as investment managers, and packaged junk assets to unsuspecting policyholders—this is the captive insurance scheme. To conceal the fraud, regulations in states like Vermont allow original insurers and affiliated reinsurers to privately set up reinsurance risk assets with arbitrarily sized capital buffers.
Take Terra Luna as an example: if Do Kwon had access to PE resources, he might have acquired an insurer called Alameda Insurance, using premium funds to buy USDT and shore up the peg. Alameda sells life insurance to Californians, holding billions in funds and purchasing investment-grade corporate bonds. Luna bribes Moody's (MCO.US) analysts to rate its corporate debt investment-grade, offering yields 5% above the 10-year Treasury. Alameda registers a reinsurer in Vermont called Three Daggers, pledging only $1 of equity for every $100 in reinsurance assets.
Once USDT's price drops and Luna can't meet bond interest payments, ratings get downgraded, the entire structure collapses, Alameda becomes technically insolvent, and policyholders suffer enormous losses. Most US state insurance guarantee limits top out at just $250,000 to $300,000, leaving the shortfall unrecoverable. Moreover, guarantee funds are funded ex-post by surviving insurers, which perversely incentivizes extreme risk-taking.
History rhyming suggests bailout inevitability and shapes the crypto outlook. In 2008, AIG (AIG.US) acted as the ultimate absorber of toxic subprime CDO squared debt, and the government stepped in to rescue it. After TARP funds plugged the AIG hole, the money flowed straight to Goldman Sachs (GS.US), which paid record bonuses in 2009. Paulson and Bernanke had held the line back then by letting Lehman fail, but that decision was a mistake that exposed how financial institutions plunder the public.
This time, Bessent and Warsh will never allow a major insurer to go bankrupt, replaying a financial disaster movie like The Big Short. They will keep printing money to avert a reckoning because, after 2008, populist political forces have risen, and the public won't be as compliant as before. When Obama approved the bailout, he didn't launch mass foreclosure prevention. By 2028, AOC won't be so accommodating. Therefore, Warsh and Bessent must avoid a public credit catastrophe at all costs.
If Trump declines to be the final purchaser of computing power and rating agencies downgrade AI data center debt, money printing will roll out in measured, gradual doses to prevent the market from fully realizing the insurance industry's insolvency. As founder of the AI/crypto project Flop Network, Hayes believes this macro environment is highly favorable. The US government won't let free markets halt data center construction; raw computing costs will decline, spot supply will glut, and AI agent adoption will accelerate.
Beyond that, the flood of new dollars will drive capital into crypto assets, which historically perform strongest during monetary expansion cycles.
Comments