The AI Boom's Profits Sit Furthest From the Customer
Author: Protik Ganguly
Morgan Stanley estimates that hyperscalers and AI cloud providers will spend about $980 billion in 2026 and $1.58 trillion in 2027, and that nearly $700 billion of the roughly $1 trillion of 2027 spending landing in S&P 500 revenue goes to one company, Nvidia (Fortune, 2026). The same report counts 91.7% of Nvidia's 2027 revenue as exposed to AI infrastructure demand. That is an exposure measure, not a share of sales, but it shows where the money lands first.
The promise of the boom was that AI would multiply what every business can produce. The drift is in the timing. A chip supplier is paid when the order is placed. The customer is paid, if at all, over the years the hardware runs. Demand is real: Microsoft's AI business reached a $37 billion annual run rate, up 123% in a year (Kindig, 2026). But the spending on the other side is larger and arrives earlier, so the supplier collects before the customer knows whether the purchase pays.
We have seen this arrangement before. In 2000, Lucent committed $5.7 billion in credit to telecom customers so they could buy its equipment, a practice called vendor financing. Winstar, which Lucent had lent $2 billion, later filed for bankruptcy. Cisco and Nortel had made similar loans (eWeek, 2001). When the customers failed, the suppliers' sales and their loan books failed together.
Today's version differs in one important way. The biggest customers are among the most cash-generative companies ever to exist, and this is not a prediction that they will fail. The point is narrower. A supplier's profit depends on its customers' returns arriving eventually, and those returns depend on revenue that has to grow into spending of this size. Morgan Stanley itself flags that 2027 should bring the largest cash-flow deficit of its forecast period, that part of the spending growth reflects higher component prices rather than more capacity, and that a plateau would lower prices while volumes hold (Fortune, 2026).
The useful habit is to ask two questions of any boom. Who is paid first? And what does that payer depend on? When the answer to the second is the customer's future profit, the supplier's results are really a measure of the customer's confidence. That does not make the supplier a bad business. It makes its fortunes tied to a payment that has not yet been earned.
What nobody knows is whether AI revenue grows into the spending before financing tightens. This is a view of how the money flows, not a forecast.
References
Lichtenberg, N. (2026, October 19). Morgan Stanley's favorite AI stock is 92% exposed to the $1.6 trillion buildout it's bullish on. Fortune. https://fortune.com/2026/10/19/nvidia-concentration-hyperscaler-1-6-trillion-92-percent-exposure/
Kindig, B. (2026, August 6). Big Tech's AI revenue is surging, but suppliers will still be the bigger winners.
Related Articles
The Bond Market Is Financing the AI Boom. Does It Know the Full Risk?
• AI‑related global debt issuance surged to $236 billion through May 2026, projected to reach $570 billion by year‑end, with major tech firms issuing $159 billion in corporate bonds in the first five months—far exceeding their borrowing in the prior five years. • Bond investors have treated AI infrastructure debt as low‑risk, but the rapid expansion of data‑center and chip financing has created a potential maturity mismatch, where the value of physical AI assets could decline before the long‑dated debt is repaid. • Regulatory uncertainty is growing, highlighted by the proposed Sanders‑AOC AI Data Center Moratorium Act and multiple state and local actions that have delayed or blocked projects, prompting early signs of buyer caution in the bond market.
The Uncomfortable Middle — What the Transition Actually Costs
• The transition to the Industrial Revolution was marked by the Luddite uprisings, where textile workers destroyed machinery due to desperation about their skills being made obsolete. This event highlights the gap between disappearing jobs and new, yet-to-be-invented ones. • The current AI transition is expected to replace many jobs, with computer scientist Geoffrey Hinton predicting significant capabilities in 2026. AI systems are rapidly completing tasks that previously took longer, with a recent MIT analysis finding they can handle nearly 12% of tasks tied to the US labor market. • Entry-level white-collar roles are being affected first by AI, with tasks such as drafting, research, and analysis being handled by machines, potentially hollowing out junior positions that once performed these tasks.
AI Is Making Tasks Faster. It Isn't Making Companies Richer.
• American companies are expected to spend $675 billion on AI infrastructure in 2026, a 63% increase from last year, but the financial returns on this investment are unclear. • Studies have shown that AI can make tasks faster and improve output quality, but this does not necessarily translate to increased company-level productivity. • Only a small percentage of companies, such as 21% of S&P 500 companies, can cite a measurable AI benefit, and many executives struggle to confidently measure AI return on investment.