Artificial intelligence is now a capital markets story
But AI is no longer solely a story about models, applications, and productivity. It is increasingly about physical infrastructure: chips, data centres, power, grid connections, land, and construction.When a technology theme starts competing for infrastructure, it progresses from a software story to a capital cycle – one that creates value but demands disciplined financing and clear returns.
The numbers are significant. Analysts forecast that the five largest US hyperscalers could spend more than $4.5 trillion in capital expenditures (capex) over the next five years, with roughly $3.2 trillion financed through capital markets, according to Bloomberg consensus estimates.
The spending is front-loaded. Companies are investing today to secure capacity, scale advantages, and market share, while productivity gains and monetisation are expected to build over time. That timing gap is why capital markets are becoming central to the AI story.
The value of investments, and the income from them, can fall as well as rise and you may not get back what you put in. Past performance should not be taken as a guide to future performance. You should continue to hold cash for your short-term needs. This article should not be taken as advice.
Capex is absorbing more free cash flow
- The first phase of the AI build-out was largely funded from existing operations.
- The leading hyperscalers are still highly profitable, and many continue to have strong financing flexibility. However, capex intensity has risen sharply.
- Free cash flow has not disappeared. Rather, a large share of it is being redirected towards AI infrastructure.
- Investors should pay closer attention to the funding mix, borrowing costs and returns on invested capital.
AI is becoming an infrastructure story
- A growing proportion of capex is dedicated to physical infrastructure, with grid connection delays and other supply chain constraints slowing deployment.
- Infrastructure bottle necks are changing not only what is being built, but when returns are realised.
- Infrastructure projects typically have longer lead times, higher upfront costs and more gradual revenue realisation, relative to developing software and applications.
- AI can support productivity growth over time. But the market is now financing the infrastructure required for those future gains, and that raises the importance of execution, utilisation and cash-flow visibility.
Credit markets are becoming part of the theme
- Technology has become a larger share of both investment grade and high yield credit markets, with sector weightings and issuance volumes rising over the past year.
- Off-balance sheet financing structures have also expanded rapidly as companies benefit from funding while preserving balance sheet flexibility. For bond investors, this adds complexity, making it harder to gauge the true scale and timing of obligations.
- This broadening is a positive feature of the cycle. It reduces reliance on any single funding channel and helps explain why markets have successfully absorbed supply so far.
- But valuation still matters. In our view, current spreads do not fully compensate for the uncertainty around timing, execution and utilisation.
Equities and credit: different implications
AI capex has different implications for equities and credit.
For equities, the theme remains attractive. Equity investors participate in the upside if AI adoption accelerates, revenues grow, margins expand and scarce assets, such as compute, data, and power capacity, command higher economic rents.
Credit investors receive coupon and principal. At current spreads, they do not participate meaningfully in the upside if AI monetisation is stronger than expected. But they are exposed to the downside if free cash flow weakens, refinancing conditions tighten, projects are delayed or utilisation disappoints.
That asymmetry is not a reason to avoid AI-related credit altogether. Many large issuers remain fundamentally strong and have managed their financing carefully. It does, however, argue for selectivity.
A powerful theme, but not a free lunch
In short, the next phase requires investors to look beyond the growth story and assess how efficiently capital is deployed, financed, and converted into cash flow. AI should be analysed both as a technology theme and a capital cycle.
The views expressed in this Sponsor article are the author's own and do not necessarily represent those of Cambridge Tech Week.