Funding the Next Dollar: AI and the Marginal Supply of Capital

funding-next-dollar

On the surface, financial conditions appear to be easing: the FOMC held rates steady last week, and Treasury yields have retreated from their recent highs. But that rally reflects a flight to safety, driven by growing questions about the durability and financing of the AI investment cycle.

As we mentioned in last week’s Notes, falling Treasury yields are only one part of the puzzle; other measures of financial conditions remain tight. Equity valuations and corporate spreads shape the private sector’s ability to tap capital markets, while volatility weighs on risk appetite. And although Chair Warsh has expressed disdain for forward guidance, his rhetoric has nonetheless shaped expectations. The FOMC does not necessarily need to raise rates for financial conditions to tighten, particularly as sentiment around AI shifts and capital becomes less readily available in parts of the market.

For corporate borrowers, the effective cost of debt is roughly the underlying government bond yield plus the credit spread investors demand to take on corporate risk. If Treasury yields fall while spreads widen, the benefit of a lower risk-free rate can be partially or fully offset. There are signs of a “crowding out” effect, in which a critical mass of borrowers competes for a finite pool of investor capital, potentially raising financing costs for everyone. Today, that competition is coming from both the public and private sectors. The U.S. government continues to issue large amounts of Treasury debt to finance persistent deficits, while corporations are increasingly tapping debt markets to fund investment, including the enormous capital requirements of the AI buildout. Private credit adds another layer of demand for capital. Notably, hyperscalers, which have emerged as a prime source of corporate bond supply, have seen their credit spreads widen.

long-investment

See more: When Protecting Benefits Gets in the Way of Building Independence

The two largest AI labs offer an interesting case study. In June, Kroll assigned Anthropic a BB+ rating, placing it below the investment grade threshold, and Blackstone reportedly anticipated ratings in the BB range following an IPO. Separately, one Anthropic chip financing transaction was reportedly priced at a yield above 10%. Raising debt capital as a high-yield issuer presents its own challenges, given that the high-yield market is a fraction of the size of the investment grade bond market. OpenAI faces a similar constraint in the equity market: it was recently reported that the company reduced its latest capital raise from roughly $30 billion to $20 billion and lowered its proposed valuation from $1.40 trillion to $1.25 trillion. For an investment cycle as capital intensive as the current AI boom, the shift from abundant to selective capital may ultimately matter more than a 25-basis-point change in the policy rate.