Funding the Future of AI: A $3 Trillion Challenge

Joining the AI race also requires building out massive physical infrastructure. Our Head of Corporate Credit Research Andrew Sheets explains why credit markets may play a critical role in the endeavor.

Read more insights from Morgan Stanley.

----- Transcript -----

Andrew Sheets: Welcome to Thoughts on the Market. I'm Andrew Sheets, Head of Corporate Credit Research at Morgan Stanley.

Today – how the world may fund $3 trillion of expected spending on AI. 

It's Friday July 25th at 2pm in London.

Whether you factor it in or not, AI is rapidly becoming a regular part of our daily lives. Checking the weather before you step out of the house. Using your smartphone to navigate to your next destination, with real time traffic updates. Writing that last minute wedding speech. An app that reminds you to take your medication or maybe reminds you to power off your device.

All of these capabilities require enormous physical infrastructure, from chips to data centers, to the electricity to power it all. And however large AI is seen so far, we really haven't seen anything yet. Over the next five years, we think that global data center capacity increases by a factor of six times. The cost of this spending is set to be extraordinary. $3 trillion by the end of 2028 on just the data centers and their hardware alone. 

Where will all this money come from? 

In a recent deep dive report published last week, a number of teams within Morgan Stanley Research attempted to answer just that. First, large cap technology companies, which are also commonly called the hyperscalers. Well, they are large and profitable. We think they may fund half of the spending out of their own cash flows. 

But that leaves the other half to come from outside sources. And we think that credit markets – corporate bonds, securitized credit, asset-backed finance markets – they're gonna have a large role to play, given the enormous sums involved.

For corporate bonds, the asset class closest to my heart, we estimate an additional $200 billion of issuance to fund these endeavors. Technology companies do currently borrow less than other sectors relative to their cash flow, and so we're starting from a relatively good place if you want to be borrowing more – given that they're a small part of the current bond market. While technology is over 30 percent of the S&P 500 Equity Index, it's just 10 percent of the Investment Grade Bond Index.

Indeed, a relevant question might be why these companies don't end up borrowing more through corporate bonds, given this relatively good starting position. 

Well, some of this we think is capacity. The largest non-financial issuers of bonds today have at most $80 to $90 billion of bonds outstanding. And so as good as these big tech businesses are, asking investors to make them the largest part of the bond market effectively overnight is going to be difficult. 

Some of our thinking is also driven by corporate finance. We are still in the early stages of this AI build out where the risks are the highest. And so, rather than take these risks on their own balance sheet, we think many tech companies may prefer partnerships that cost a bit more but provide a lot more flexibility. One such partnership that you'll likely to hear a lot more about is Asset Backed Finance or ABF. We see major growth in this area, and we think it may ultimately provide roughly $800 billion of the required funding.

The stakes of this AI build out are high. It's not hyperbole to say that many large tech companies see this race to develop AI technology as non-negotiable. The cost of simply competing in this race, let alone winning it – could be enormous. The positive side of this whole story is that we're in the early innings of one of the next great runs of productive capital investment, something that credit markets have helped fund for hundreds of years. 

The risks, as can often be the case with large spending, is that more is built than needed; that technology does change, or that more mundane issues like there not being enough electricity change the economics of the endeavor.

AI will be a theme set to dominate the investment debate for years to come. Credit may not be the main vector of the story. But it's certainly a critical part of it. 

Thank you as always, for your time. If you find Thoughts on the Market useful, let us know by leaving a review wherever you listen. And also tell a friend or colleague about us today.