Amazon has joined Microsoft in pausing some datacenter leasing deals, sparking fresh concerns about whether the AI hype train may be running out of steam.

According to a note issued by financial services biz Wells Fargo, the ecommerce giant is taking a breather from some of its negotiations over leasing of colocation capacity for its Amazon Web Services (AWS) cloud division, particularly those involving international partners.

The note states that AWS is not canceling already signed agreements, but merely pulling back from discussions over additional capacity in the pipeline.

Amazon itself says there’s nothing to see here – it’s merely business as usual, but it follows soon after fellow hyperscaler Microsoft also started drawing back from leasing deals, having found itself “in an oversupply position,” and there is growing unease over the global economic situation caused by the Trump administration’s quixotic approach to trade.

In a LinkedIn post in response to the news, AWS VP for Global Datacenters Kevin Miller wrote: “There continues to be significant interest and speculation about the datacenter expansion plans of AWS and other cloud providers.”

Miller claimed that there is no fall-off in demand for AI-related cloud services, and that the moves are simply a result of the retail giant’s cloudy arm seeking the best deals to meet its capacity requirements.

“First and foremost, we continue to see strong demand for both Generative AI and foundational workloads on AWS. We have almost two decades of experience delivering datacenter capacity to meet customer demands, when and where they need it. That experience has taught us to consider multiple solutions in parallel. Some options might end up costing too much, while others might not deliver when we need the capacity. Other times, we find that we need more capacity in one location and less in another.

“This is routine capacity management, and there haven’t been any recent fundamental changes in our expansion plans,” Miller concluded.

This was echoed by Vlad Galabov, Senior Research Director for Enterprise Infrastructure at Omdia, who told The Register that many investors do not understand the compute requirements for the current wave of AI expansion.

“My take is – investors are clutching at straws to find a fault in the datacenter industry,” Galabov said. “The fact is we need and will continue to need more compute. AI is one driver for this insatiable demand, but the continued transition towards a more and more digital global economy, and the slow ramp of specialized computing all require more compute power.”

“It is normal for hyperscale cloud service providers to right-size their colocation capacity, given that they also build their own datacenters. This has always happened,” he added, chiding observers for examining every decision that is made with a fine-tooth comb, looking for signs that a supposed AI bubble is bursting.

“I spoke to a lot of investors at the Omdia Analyst Summit at Datacenter World, and the commonality was they did not understand the profound need for computing until we went deep into the computing requirements of both AI training and inference,” Galabov said.

However, Fabrice Coquio, SVP of global colocation biz Digital Realty had a slightly different take, saying there could be an AI bubble precisely because investors don’t understand it.

“Is there a bubble? Potentially? I see the risk, because when some of the traditional investments in real estate – like housing, logistics, and so on – are not that important, people are looking to invest their amazing capacity of available funds in new segments, and they say, ‘Oh, why not datacenters?'” he said in an interview with The Register recently.

“In the UK, in France, in Germany, you’ve got people coming from nowhere having no experiences … that have no idea about what AI and datacenters are really and still investing in them,” he added.

At the moment, about 20 organizations represent roughly 60 percent of capacity bookings around the world, Coquio said.

“Most of them have completed their time to market initiative to be in multiple areas where they needed multiple subcontractors for colocation services. They are now engaged for many years already in reducing the number of partners, either because they build more and more themselves, or when they need absolutely a collaboration provider, providing connectivity, concentration of customers and so on, they reduce that to the bare minimum,” he told us.

So it isn’t a sign of the AI bubble popping just yet, though we note the growing unease in the markets at the Trump administration’s trade policies, and that investment in AI bit barn capacity continues to heavily outweigh the amount enterprise customers are spending on AI software licenses. ®