OpenAI's CFO using the term "backstop" doomed the request by associating AI investment with the 2008 bank bailouts. The word conjures failure and socializing private losses, whereas a term like "partnership" would have framed the government's role as a positive, collaborative effort, avoiding immediate public opposition.
Related Insights
The call for a "federal backstop" isn't about saving a failing company, but de-risking loans for data centers filled with expensive GPUs that quickly become obsolete. Unlike durable infrastructure like railroads, the short shelf-life of chips makes lenders hesitant without government guarantees on the financing.
OpenAI's CFO hinted at needing government guarantees for its massive data center build-out, sparking fears of an AI bubble and a "too big to fail" scenario. This reveals the immense financial risk and growing economic dependence the U.S. is developing on a few key AI labs.
New technologies perceived as job-destroying, like AI, face significant public and regulatory risk. A powerful defense is to make the general public owners of the technology. When people have a financial stake in a technology's success, they are far more likely to defend it than fight against it.
Acknowledging a de facto government backstop before a crisis encourages risky behavior. Lenders, knowing their downside is protected on AI infrastructure loans, are incentivized to lend as much as possible without proper diligence. This creates a larger systemic risk and privatizes profits while socializing eventual losses.
AI is experiencing a political backlash from day one, unlike social media's long "honeymoon" period. This is largely self-inflicted, as industry leaders like Sam Altman have used apocalyptic, "it might kill everyone" rhetoric as a marketing tool, creating widespread fear before the benefits are fully realized.
The public is unlikely to approve government guarantees for private AI data centers amid economic hardship. A more palatable strategy is investing in energy infrastructure. This move benefits all citizens with potentially lower power bills while still providing the necessary resources for the AI industry's growth.
Drawing from the nuclear energy insurance model, the private market cannot effectively insure against massive AI tail risks. A better model involves the government capping liability (e.g., above $15B), creating a backstop that allows a private insurance market to flourish and provide crucial governance for more common risks.
Following backlash over his CFO's comments, Sam Altman reframed the request away from government guarantees for private companies. Instead, he proposed the government build and own its own AI infrastructure. This strategically repositions the ask as creating a public asset where financial upside flows back to the government.
The current market boom, largely driven by AI enthusiasm, provides critical political cover for the Trump administration. An AI market downturn would severely weaken his political standing. This creates an incentive for the administration to take extraordinary measures, like using government funds to backstop private AI companies, to prevent a collapse.
OpenAI publicly disavows government guarantees while its official documents request them. This isn't hypocrisy but a fulfillment of fiduciary duty to shareholders: securing every possible advantage, including taxpayer-funded incentives, is a rational, albeit optically poor, corporate best practice.
