During the early part of my career in Baltimore, I managed a couple of money market common trust funds, among other portfolios. These invested in repurchase agreements, corporate commercial paper, bankers’ acceptances, federal agency discount notes, U.S. Treasury bills and other debt securities that matured within 397 days.
At the time, 1992 to 1994 or thereabouts, the wizards of Wall Street were creating a number of somewhat complicated money market investments. These included inverse floaters, dual-index floaters and other securities that required way too much analysis to determine how they worked. (1)
Intuitively, that isn’t what you really want in a money market security.
One day, and I distinctly remember it, a representative from a brokerage firm in Texas tried extremely hard to sell me several million dollars’ worth of some crazy product I couldn’t grasp. I told him so and asked why it had to be so complicated. To that, he admitted he didn’t fully understand it either and would “get his trader” on the phone to explain it.
Not surprisingly, or so I would hope, I told him to forget it. We weren’t going to buy something we didn’t understand. The thought process is simple. The more complex an investment is, the more money the person who created it will undoubtedly make. In this case, the more money they stood to make on this instrument, the less money our clients stood to make, by definition, really.
If you are a history buff, you might remember many of these securities effectively blew up when the Fed started raising the overnight rate in 1994. At the time, the media referred to them as “exotic derivatives.” Our firm just thought of them as unnecessary. (2)
Suffice it to say, it ended up being a real mess, one that we didn’t have to clean up, but a lot of other fund companies did.
This has been a rule of thumb for me ever since: don’t buy what you don’t understand.
It isn’t terribly complicated, and it doesn’t mean I don’t like risk. Far from it. In fact, at times, following this self-imposed regulation can be extremely risky in terms of relative performance. However, over the long run, it has generally served me well.
With this in mind, I am currently concerned about the sheer scope of the artificial intelligence (AI) infrastructure buildout. Don’t get me wrong. I don’t think there is any debate AI will fundamentally change our lives and how we conduct business. It will. Further, the recent rapid advancements in technology are both mind-numbing and really cool.
However, will these investments actually be profitable? Are they worth the massive expenditure of money? Those are far different questions than asking whether the technology is life-changing.
I am increasingly having trouble understanding how these investments will turn a profit. The math doesn’t seem to add up.
By some estimates, the total investment in the AI buildout could exceed $10 trillion by 2032. That is $10 trillion with a T, and that is a lot of money. If we assume companies are able to finance the entire $10 trillion at an annual interest rate of 7%, this works out to be $700 billion in annual interest payments alone. Further, if we assume a straight-line 10-year amortization schedule, that represents an additional $1 trillion per year. Combined, somewhat obviously, that is $1.7 trillion in first-year debt service alone. (3)
To put that number into perspective, it exceeds the IMF’s April 2026 estimates for the annual economic output of both Turkey and Indonesia. Think about that for a second. In order to potentially service that amount of accumulated debt in 2032, these companies would need to generate cash equivalent to basically the entire economic output of Turkey. (4)
I am not talking about operating profit margins, COGS, SG&A or any of that. Just the necessary cash to keep from defaulting. As such, the amount of revenue these companies will have to generate on their AI investments to turn a profit is some number well in excess of $1.7 trillion.
The knee-jerk reaction might be: “Norris, they probably aren’t going to finance that entire amount, let alone all of it at 7%.” Fair enough. However, in business school, my professors taught me equity was the most expensive form of capital. If so, adding more equity to the financing mix would make the AI buildout even more expensive than it already is. (5)
Of course, equity does not require the same out-of-pocket cash payments that debt might. However, that doesn’t mean equity is free. Far from it. Businesses need to see a return on it.
But what amount of revenue must these companies generate in order to make a profit?
Not surprisingly, this will be company-dependent, since each will have its own balance sheet, income statement and financing mix. However, what if we assume the industry would like to keep its net interest expense at 25% of operating profit or less while maintaining, say, a 25% operating profit margin (OPM)? What would the math look like?
Here goes nothing: $700 billion divided by 25% equals $2.8 trillion. This would be the required annual operating profit. Divide that $2.8 trillion by 25%, and you get $11.2 trillion in revenue. Voila. That is how much top-line revenue these projects would need to generate under these assumptions to keep net interest expense at 25% of operating profit and maintain a 25% OPM.
Obviously, that doesn’t include the amortization schedule I mentioned earlier. Just the interest payments.
Admittedly, there is “more to it” than that, and my math is simplistic. There are all kinds of incentives and other factors that might drive down costs to some degree. That is a job for the CFO’s office. No argument. However, these calculations should give you a basic idea of how much money we are discussing here. We are talking about trillions upon trillions of dollars.
Is this possible? To be sure, it might be. Only time will tell. But is it likely or probable? Tell me what your definitions of those words are. Also, give me some odds.
In the end, given even odds, I think it will be difficult for the industry to generate enough revenue to maintain margins and continue to invest in AI the way it currently is. I am simply not sure I understand the math and growth assumptions behind these investments.
This takes me back to my original rule of thumb.
Thank you for your continued support. As always, I hope this newsletter finds you and your family well. May your blessings outweigh your sorrows on this and every day. Also, please be sure to tune into our podcast, Trading Perspectives, which is available on every platform.
John Norris
Chief Economist
SOURCES:
- The Edge Malaysia – Derivatives World: The Misadventures of Derivatives: The 1990s. April 19-25, 2010.
- Journal of Accountancy – Derivatives Revisited. May 1, 2000.
- The Wall Street Journal – The AI Build-Out Is Becoming the Biggest Economic Bet in U.S. History. Sept. 24, 2026.
- Wikipedia – List of Countries by GDP (Nominal).
- Investopedia – Cost of Debt vs. Equity: Key Differences and Impact on Business. Updated Feb. 18, 2026.
About Common Cents:
Please note, nothing in this newsletter should be considered or otherwise construed as an offer to buy or sell investment services or securities of any type. Any individual action you might take from reading this newsletter is at your own risk. My opinion, as well as those of our Investment Committee, is subject to change without notice. The opinions expressed herein are not necessarily those of the rest of the associates and/or shareholders of Oakworth Capital Bank or the official position of the company itself.
Opinions and forward-looking statements are subject to change without notice and may not come to pass. Past performance is not indicative of future results. References to market indices are for illustrative purposes only. Index performance does not reflect the deduction of fees or expenses, and individuals cannot invest directly in an index. Political and geopolitical events are referenced for general context and do not represent a recommendation or prediction regarding any specific investment outcome. Links to third party articles are provided for informational purposes only. Oakworth Asset Management LLC does not control, endorse or adopt the content, opinions or recommendations expressed on third party websites. Such content may be dated and may not reflect current market conditions. Third party information is believed to be reliable but has not been independently verified. Readers should evaluate the information independently. Any reference to specific securities is provided for informational and illustrative purposes only and should not be considered a recommendation to buy, sell, or hold any particular investment. It should not be assumed that any such securities discussed were or will be profitable.
About Oakworth:
Oakworth Capital Bank, Member FDIC, Equal Housing Lender, provides commercial and private banking, wealth management and advisory services. Advisory Services, including investment management and financial planning, are offered through Oakworth Asset Management LLC, a registered investment advisor. Oakworth Asset Management LLC is owned by Oakworth Capital Bank. Investment products and services offered via Oakworth Asset Management LLC are independent of the products and services offered by Oakworth Capital Bank, and are not FDIC insured, may lose value, have no bank guarantee, and are not insured by any federal or state government agency. Because of the ownership relationship and involvement by Oakworth Asset Management LLC associates with Oakworth Capital Bank, there exists a conflict of interest to the extent that either party recommends the services of the other. For additional information about Oakworth Asset Management LLC, including its services and fees, visit adviserinfo.sec.gov.