Meta Will Run its Servers For Up to 5 Years

Feb. 2, 2023
Meta confirmed that it is canceling multiple data center construction projects, and is also extending the expected life span of its servers to 5 years. Both moves are designed to optimize Meta's investment in its massive data infrastructure.

Meta Platforms will build its data centers faster and run its servers for longer. Both moves are part of the company's strategy to be more efficient with its investments in its massive global data infrastructure. 

In Wednesday's earnings announcement, Meta confirmed that it is canceling multiple data center construction projects, while shifting others to a new design that should be faster and cheaper to build. 

The company is also extending the expected life span of its servers and network equipment, a move which will boost Meta's bottom line by about $1.5 billion for 2022. In December Meta determined that it can now use most of its servers and network devices for five years, up from 4 years as of year-end 2021.

Both moves will help Meta save money as its faces a difficult economy and the evolving landscape for virtual reality and generative AI (artificial intelligence), two technologies that will guide its future. The company reduced its projection for 2023 capital expenses by about $4 billion, which reflects both the abandoned data center projects and a consolidation of Meta's office leases.  

Building Fewer Data Centers Faster 

In December, Meta decided to overhaul its data center design to optimize its facilities for artificial intelligence, while also pausing construction on a number of its data center projects so it can "rescope" its new data centers with the pending design focused on AI,

Meta has not publicly itemized the data center projects that will be halted, but previous media reports indicate that it will cancel the planned expansion of its campus in Odense, Denmark, and pause recently-announced expansions in Texas, Alabama and Idaho. DCF has reached out to Meta for additional details of its plans.  

"We have streamlined our future data centers to a new architecture, which we believe will be more cost efficient and more flexible that provides us optionality to support both AI and non-AI workloads," said Susan Li, the Chief Financial Officer at Meta, in Wednesday's earnings call. "We’re expecting that the new design will be cheaper and faster to build than previous data center architecture. We have a new phased approach that allows us to build base plans with less initial capacity and less initial capital outlay, but then flex up future capacity quickly if needed. We’re still planning to grow AI capacity significantly."

 Meta operates 21 data center campuses around the globe to support its Facebook, Instagram and Messenger services. These data centers represent an investment of $16 billion and more than that span 40 million square feet of space. Meta typically builds between five and nine data center buildings on each of its huge campuses. 

Servers Are Sturdier Than We Thought

Meta had previously extended the expected life of its servers in the second quarter of 2022, changing from 4 years to 4.5 years. Just two quarters later, the company further extended its estimate to 5 years. The financial impact of the changes was a reduction in depreciation expense of $860 million and an increase in net income of $693 million, Meta said. 

The depreciation change only extends the useful lives of non-AI servers. With the rapid innovation in artificial intelligence algorithms and learning models, AI hardware will need to be updated more frequently than general purpose servers. 

Meta joins the largest cloud platforms in extending its estimates on server lifeAs of 2020, the estimated lifespan of a server on a cloud computing platform was three years, an assumption embraced by all three of the leading platforms. Now Amazon Web Services now says its servers have a ‘useful life” of five years, while Google and Microsoft expect servers to last for four years.

These estimates are developed for accounting purposes, and can shift billions of dollars to a company’s bottom line by deferring spending on hardware and enabling assets to be expensed over a longer period. But the trend toward more durable cloud servers is also likely influenced by design refinements, and may help manage server fleets differently to support continued growth during supply chain challenges in the semiconductor industry.

Longer server life will also matter in navigating the supply chain challenges in the chip sector, which are expected to persist into the second half of 2022. The limited availability of new chips has created delivery challenges for many industries including automakers and equipment vendors.

About the Author

Rich Miller

I write about the places where the Internet lives, telling the story of data centers and the people who build them. I founded Data Center Knowledge, the data center industry's leading news site. Now I'm exploring the future of cloud computing at Data Center Frontier.

Sponsored Recommendations

The AI Disruption: Challenges and Guidance for Data Center Design

From large training clusters to small edge inference servers, AI is becoming a larger percentage of data center workloads. Learn more.

A better approach to boost data center capacity – Supply capacity agreements

Explore a transformative approach to data center capacity planning with insights on supply capacity agreements, addressing the impact of COVID-19, the AI race, and the evolving...

How Modernizing Aging Data Center Infrastructure Improves Sustainability

Explore the path to improved sustainability in data centers by modernizing aging infrastructure, uncovering challenges, three effective approaches, and specific examples outlined...

How Modern DCIM Helps Multi-Tenant Colocation Data Centers Be More Competitive

Discover the transformative impact of modern DCIM software on multi-tenant colocation data centers, enhancing competitiveness through improved resiliency, security, environmental...

Sashkin/Shutterstock.com

Unpacking CDU Motors: It’s Not Just About Redundancy

Matt Archibald, Director of Technical Architecture at nVent, explores methods for controlling coolant distribution units (CDU), the "heart" of the liquid cooling system.

White Papers

Dcf Imdcwp Cover 2023 01 11 17 19 43

Infrastructure Planning Report - EMEA - Frankfurt

Jan. 11, 2023
In this white paper, Iron Mountain Data Centers provides an overview of the German colocation market. It explores strengths and weaknesses of the market as well as the latest ...