AI shifts from tech to capital markets. $1.5T+ financing needs through 2028 push debt into utilities, infrastructure, and private credit.
Watch time: {{ formattedDuration }}
View transcript
Good afternoon and welcome back to The Weekly Fix.
Last week’s stronger-than-expected jobs report pushed Treasury yields higher and led investors to further scale back expectations for Fed rate cuts. While markets continue debating the near-term path of monetary policy, however, another development is emerging that may end up mattering more over the long run.
Most investors are aware of the growing number of technology companies tapping capital markets at unprecedented scale. What is less clear is whether this indicates AI is entering a new phase, one where access to capital becomes almost as important as access to technology.
Last week, Alphabet announced plans to raise approximately $85 billion of equity capital while simultaneously guiding to roughly $180 to $190 billion of capital expenditures for 2026 and even higher spending in 2027. At the same time, Anthropic confidentially filed for an IPO, while SpaceX is preparing what could become one of the largest public offerings in history.
Viewed independently, these are company-specific events. Viewed together, they suggest that AI may be shifting from innovation and technology dependence to capital and financing dependence.
Historically, the most important technology revolutions have eventually become capital formation stories. Like railroads in the 1800s or fiber networks in the 1990s, AI may be entering a similar phase.
Industry estimates show that an additional $1.5 trillion or more of AI-related infrastructure financing will be required through 2028—an amount approaching the size of the entire U.S. leveraged loan market. That figure includes data centers, power generation, transmission infrastructure, networking equipment, and other assets required to support the next generation of compute capacity. This is in addition to the roughly $1.5T of internally generated capital expected from the companies themselves.
The important implication is that the financing requirement is increasingly larger than what hyperscaler balance sheets alone may comfortably absorb.
The likely result is a migration of the financial burden, not just into investment-grade technology bonds, but into utility debt, infrastructure finance, project finance, private credit, securitized credit markets, and as we are seeing now, public equity markets.
This may help explain why one of the world’s most profitable companies is choosing to raise equity despite already generating enormous operating cash flow.
The question for markets and specifically fixed income investors then becomes how much leverage can the system take.
On the one hand, AI may ultimately resemble cloud computing rather than telecom. If monetization arrives quickly enough, hyperscaler cash flows could grow fast enough to fund much of the required investment internally, reducing the need for external capital and allowing for the orderly repayment of debt loads.
On the other hand, a massive injection of debt could lead to an over built or an overleveraged ecosystem.
Either way, the scale of planned spending is becoming difficult to ignore.
So while, the market ended last week debating whether rates stay higher for longer, undoubtedly and important and related discussion, the bigger question may be whether we are witnessing the early stages of a new capital markets regime—one where AI becomes less of an equity software story and more of an infrastructure debt story.
Thanks for joining, and we will see you next week.
Key takeaways
AI is becoming a capital formation story, not just a tech story: Like railroads and fiber networks before it, AI is entering a phase where access to capital matters as much as access to technology. Alphabet's $85B equity raise and SpaceX's upcoming IPO signal this transition.
The financing gap is too large for hyperscaler balance sheets alone: Industry estimates show $1.5T+ in AI infrastructure financing needed through 2028 (data centers, power, networking) will force capital into investment-grade tech bonds, utilities, infrastructure finance, private credit, and securitized markets.
The system faces a leverage inflection point: Either AI monetization arrives fast enough to fund investment internally (cloud computing scenario), or massive debt loads create an overleveraged ecosystem. Either way, the scale of spending is impossible to ignore.
