When Microsoft said in April that it would put a pause on data center construction, there was considerable hand wringing as people tried to read between the lines, even though Microsoft said flat out why it was pausing construction. That is in part that OpenAI is reportedly contemplating building its own data centers, reducing the need for further Microsoft investment. Now, one analyst thinks that the rest of the industry should follow Microsoft’s lead.
Data centers take a few years to construct coming from breaking ground to turning on the lights. In that time, technology can change radically and is doing so now. A data center constructed now could be obsolete by the time it comes online in three years. So, data center builders might be better off waiting and let the chaotic state of the market shake itself out.
“I think they’re putting the brakes on the data center expansion until the hardware reaches some form of a steady state, so when you build something, it’s not obsolete by the time you complete it,” said Rob Enderle, principal analyst with The Enderle Group. “There are just so many moving parts here it’s unclear how you would build a data center and not have it be obsolete when you’re completed.”
“Not to mention the fact that right now with the tariffs everything else going on it’s not a great time to build it out infrastructure, so you’re probably better off to expand the data centers overseas because the tariffs are going to kill you,” he added.
The Trump Administration’s tariffs, which are proving more harmful than originally thought, also serve as an incentive, if not to delay construction, then to move it overseas instead of building in the US.
“With the tariffs, stuff is way too expensive now to build domestically anyway, so it’s best to wait until the tariffs settle and if you’re going to build, build overseas, because you’re not hit by the tariff problem and you’ll get incentives” From host countries anxious to get data center business, he added.
Tariffs aside, Enderle feels that AI technology and ancillary technology around it like battery backup is still in the early stages of development and there will be significant changes coming in the next few years.
GPUs from AMD and Nvidia are the primary processors for AI, and they are derived from video game accelerators. They were never meant for use in AI processing, but they are being fine-tuned for the task. It’s better to wait to get a more mature product than something that is still in a relatively early state.
But Alan Howard, senior analyst for data center infrastructure at Omdia, disagrees and says not to wait. One reason is the rate at which people that are building data centers is all about seizing market opportunity.” You must have a certain amount of capacity to make sure that you can execute on strategies meant to capture more market share.”
The same sentiment exists on the colocation side, where there is a considerable shortage of capacity as demand outstrips supply. “To say, well, let’s wait and see if maybe we’ll be able to build a better, more efficient data center by not building anything for a couple of years. That’s just straight up not going to happen,” said Howard.
“By waiting, you’re going to miss market opportunities. And these companies are all in it to make money. And so, the almighty dollar rules,” he added.
Howard acknowledges that by the time you design and build the data center, it’s obsolete. The question is, does that mean it can’t do anything? “I mean, if you start today on a data center that’s going to be full of [Nvidia] Blackwells, and let’s say you deploy in two years when they’ve already retired Blackwell, and they’re making something completely new. Is that data center full of Blackwells useless? No, you’re just not going to get as much out of it as you would with whatever new generation that they’ve got. But then you wait to build that, then you’ll never catch you never catch up,” he said.
Source:: Network World