Meta has axed a major data center expansion in Odense, Denmark, laying off the primary contractor on two facilities in the country.
According to reporting by Data Centre Dynamics, the move come as a result of a pivot by the company towards building data centers whose technology is more suited to the demands of modern AI workloads.
“Supporting AI workloads at scale requires a different type of data center than those built to support our regular online services,” said Meta’s Nordics head of communications Peter Münster, who added that Meta is now focusing its efforts on “building a new generation of data centers”.
What’s next for Meta’s data centres?
Meta is now reworking 11 currently under-development projects for these new AI workloads, which themselves are still under development, as the new designs are yet to be finalized according to the publication’s sources.
GPUs built explicitly for AI workloads or machine learning (ML) databases can, for example, require much more electrical power to run effectively.
In addition, Meta has also announced that it is set to pause the construction of a planned 900,000 square feet data center in Temple, Texas, which the construction of only began in Spring.
Though Meta’s performance hasn’t exactly been world-beating as of late, its market capitalization has lost over half its value since January 2022, data centers are one area where it has promised to continue to invest.
Though the firm claims it is set to significantly cut expenditures in many other areas, it stated that it is set to make “ongoing investments” in its data center footprint.
A Meta spokesperson said in a call posted by Seeking Alpha: “We believe the additional data center capacity will provide us greater flexibility with the types of servers we purchase and allow us to use them for longer, which we expect to generate greater cost efficiencies over time.”
In November 2022, Meta axed around 11,000 employees, around 13% of its global workforce, saying it was “restructuring teams” to “increase efficiency”, highlighting areas such as its “real estate footprint” for cuts.
- Interested in working on your own AI projects in the cloud? Check out our list of the best cloud hosting