Contracts worth billions of dollars have become the commercial glue holding the AI boom together. But investors shouldn't rely on them sticking if the boom fades.
Contracts to supply computing power for artificial-intelligence calculations have become such a fixture of the craze that whole industries have begun reorganizing around them. AI suppliers say these arrangements give them unprecedented visibility into their future revenue, allowing them to wow investors with promises of bumper sales and profits ahead.
The computer-memory business may be the most extreme example. Memory suppliers and their customers have sought out longer-dated deals in recent months.
This is because of the explosive growth of autonomous AI agents, which are memory intensive. And that is turning what historically has been a cutthroat, cyclical business where price competition is fierce into something much more stable.
The world's big memory-chip manufacturers -- South Korea's Samsung Electronics and SK Hynix, along with U.S.-based Micron Technology -- are all raking in record profits and touting a supply shortage they expect to last into 2028. An executive at SK Hynix, which listed shares in New York this month to capitalize on the excitement about memory, suggested on a call with analysts in April that long-term deals could improve the market's view of the memory business as a whole.
Micron has been an especially active dealmaker. Its "strategic customer agreements" typically last five years and are take-or-pay. This means the buyer has to pony up whether they take shipment of the memory or not. Last month, Micron Chief Executive Sanjay Mehrotra said on an earnings call that these agreements would supply more than half of the company's revenue in the years ahead.
That sounds like a boon for investors in memory companies. And they have been strong choices of late. Micron's stock has roughly tripled this year, SK Hynix has risen nearly as much, and Samsung has about doubled.
But while the long-term contracts are helping drive the upswing, there is reason to wonder whether they provide any certainty in a downturn.
Indeed, if demand for memory should ebb before supply contracts expire, those will likely be renegotiated or extended. Chip suppliers don't like sending products to customers who aren't going to use them.
If they do, the chips will sit on the shelf until demand returns. And at that point, customers will draw down inventories before they buy new chips, delaying revenue for the chip makers.
Suppliers also don't like to force products on customers who don't want them, especially if competitors are being more flexible. Pushing unwanted memory into the supply chain could fray long-term customer relationships.
You don't have to dig too deep into tech history to find evidence suggesting that suppliers will be accommodating in a downturn.
A similar shift to longer-dated contracts happened during the Covid-era chip shortage. Chip makers used them to help fund the expansion of their manufacturing capacity and decide which customers got preferential treatment.
Microchip Technology, which makes microcontroller chips that are critical in cars and a host of electronic devices, launched a "preferred supplier program" back in 2021. When the shortage turned to a glut a few years later, the program was scrapped and customers got exemptions. Flexibility became the watchword.
Contracts were pushed out by years, and customers weren't held to taking as much supply as their Covid-era commitments stipulated. The company's chief, Steve Sanghi, put it bluntly last November: "We're not forcing customers to buy anything that they do not need."
A similar dynamic is likely to play out if the AI boom weakens. Yet the fortunes of the entire AI supply chain increasingly rest on contracts that have become larger and longer-dated as companies jostle for AI-related business.
This goes far beyond the memory market. Leading AI developers have contracts with cloud-computing firms like Oracle and CoreWeave. These players in turn have contracts to buy AI computing equipment.
The AI chip companies that supply much of that equipment have their own contracts with the likes of Taiwan Semiconductor Manufacturing to make the chips. And TSMC has long-term contracts to buy manufacturing equipment from companies like Netherlands-based ASML.
The amounts involved are often gargantuan.
Oracle, for instance, signed an enormous cloud-computing deal with OpenAI last year and had $638 billion of so-called remaining performance obligations -- essentially, contracts that the company has yet to deliver on -- at the end of its last fiscal quarter. That figure, finance chief Hilary Maxson told analysts last month, "provides exceptional visibility into our future revenue growth, all supported by long-term contractual customer commitments."
Reliance on contracts has escalated significantly in the past year. Since the middle of 2025, four big AI spenders -- Alphabet unit Google, Microsoft, Amazon.com and Oracle -- have collectively more than doubled their revenue backlogs, adding more than $1 trillion to the combined figure.
All that visibility could quickly become clouded. And the existence of big long-term contracts could worsen any AI unwind.
In its annual economic report earlier this month, the Bank for International Settlements suggested shortages across the AI supply chain could be amplifying overinvestment "as firms attempt to lock in future capacity through long-dated contracts that further expose them to any disappointments in demand."
In other words, lenders and investors who have extended funds to companies based on long-term contracts could be in for a shock.
Write to Asa Fitch at asa.fitch@wsj.com
(END) Dow Jones Newswires
July 20, 2026 05:30 ET (09:30 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments