OpenAI is reportedly contemplating building its own data center as Microsoft backs off on its own infrastructure investments.
OpenAI has privately discussed building and operating its first data center to house storage, which is essential for developing sophisticated AI models. Microsoft, on the other hand, has pulled back on its buildouts, canceling data center projects in the US and Europe.
The tech giant is also reportedly not pursuing new business with OpenAI, in which it has invested $13 billion so far. The companies’ previously exclusive partnership was modified in January, with Microsoft having right of first refusal on new capacity, and OpenAI using other cloud computing services in addition to Azure.
But these developments have broader implications for the market beyond the change in the relationship.
“It is a potentially big deal,” Alvin Nguyen, Forrester senior analyst, told Network World. If enterprises need AI services and Microsoft is pulling back its offerings, buyers will have to go to other suppliers whose terms might not be as negotiable when market options are limited.
“If a major player diminishes their footprint, and if demand is still outstripping supply, [enterprises] have to make choices that will probably end up costing more in the long run,” Nguyen pointed out.
A potential ‘oversupply position’
In a new research note, TD Cowan analysts reportedly said that Microsoft has walked away from new data center projects in the US and Europe, purportedly due to an oversupply of compute clusters that power AI.
This follows reports from TD Cowen in February that Microsoft had “cancelled leases in the US totaling a couple of hundred megawatts” of data center capacity. The researchers noted that the company’s pullback was a sign of it “potentially being in an oversupply position,” with demand forecasts lowered.
OpenAI, for its part, has reportedly discussed purchasing billions of dollars’ worth of data storage hardware and software to increase its computing power and decrease its reliance on hyperscalers. This fits with its planned Stargate Project, a $500 billion, US President Donald Trump-endorsed initiative to build out its AI infrastructure in the US over the next four years.
Based on the easing of exclusivity between the two companies, analysts say these moves aren’t surprising.
“When looking at storage in the cloud — especially as it relates to use in AI — it is incredibly expensive,” said Matt Kimball, VP and principal analyst for data center compute and storage at Moor Insights & Strategy.
“Those expenses climb even higher as the volume of storage and movement of data grows,” he pointed out. “It is only smart for any business to perform a cost analysis of whether storage is better managed in the cloud or on-prem, and moving forward in a direction that delivers the best performance, best security, and best operational efficiency at the lowest cost.”
He did question OpenAI’s move to jump into the data center business, however. “There is so much to building, powering, and managing a datacenter environment that is not typically in the DNA of a software company,” he said.
For instance, while OpenAI may be well-versed in GPUs, CPUs, and other architectures, that doesn’t necessarily translate to full-on facility management and day-to-day data center operations.
Meanwhile, it’s important to note that Microsoft still plans to invest $80 billion in data centers in this fiscal year (which ends on June 30), half of that in the US. And on the AI front, don’t count them out: “Regardless of what’s going on, I would not be surprised to see if Microsoft comes back with a vengeance in terms of AI capabilities in the future, just not OpenAI-based,” Nguyen predicted.
A market repositioning?
Some, including Alibaba Group Holding Ltd. chairman Joe Tsai, have warned of a potential bubble in data center construction following a purchasing frenzy. Others, though, say moves like Microsoft’s and OpenAI’s could signal a fundamental market restructuring.
“It’s not necessarily a slowdown; there’s still a lot of demand, and supply in terms of energy and space is still not able to meet that,” said Nguyen. “There’s going to be a lot of repositioning, who’s going to be running AI in the future?”
Microsoft’s pullback will lead to Google, Meta, AWS, and others rushing in to backfill capacity. But if there’s less market choice, costs may go up due to natural supply-and-demand cycles.
On the other hand, enterprises may opt to build data centers themselves — but GPUs are expensive, and difficult to acquire. And enterprise needs are very different from those of hyperscalers when it comes to capacity; they need a lot less, and they also want their data centers closer to their operations, Nguyen pointed out. This could lead to “more general or less powerful AI,” he said, as smaller models don’t require nearly as much equipment as highly-sophisticated models.
“This may be a potential reset moment for AI infrastructure and how much needs to be spent and how much power and cooling is required,” said Nguyen.