Musk, Zuckerberg Accelerate Data Center Spending, Reveal Competing AI Visions

Meta and Elon Musk's companies said in recent days they are amplifying their historic spending sprees on artificial intelligence data center infrastructure. Beneath the eye-popping numbers, the companies are placing sharply different bets on how the AI boom will unfold. 

The parent company of Facebook and Instagram plans to nearly double its capital spending in 2026, potentially deploying as much as $135B toward the data centers, processors and power needed to deliver gigawatts of new computing capacity, it said in its fourth-quarter earnings report.

Musk’s interwoven suite of companies — xAI, Tesla and SpaceX — is also ramping up spending on data centers and supporting equipment as Musk pivots the firms toward the creation of an integrated AI ecosystem. 

Meta and the “Muskonomy” represent a distinct cohort among the Big Tech hyperscalers driving the data center building boom.

While Amazon, Microsoft and Google are largely developing the infrastructure to support their cloud businesses, Meta and xAI aren’t building computing clusters to rent out capacity — their data center footprints exist primarily to support their own AI products.

This distinction is reflected in the similar way the two founder-led conglomerates have pursued their data center build-outs, prioritizing speed to market, scale and cheap power over considerations like latency and reliability that matter more to the cloud providers.

As a result, Meta and xAI have been the first to have gigawatt-scale data center campuses in operation. 

At the same time, the AI strategies outlined by Musk and Meta CEO Mark Zuckerberg on Q4 earnings calls diverge sharply, revealing competing visions of how the digital infrastructure landscape will evolve.

Both anticipate that many of the questions surrounding the future of AI will be answered in the months ahead, a looming inflection point requiring an unprecedented spending surge now to avoid being left behind. 

“This is going to be a very big capex year,” Musk said on Tesla’s earnings call.  “That is deliberate because we're making big investments for an epic future.”

Meta's projected capex of $115B to $135B in 2026 is well above the $72.2B it spent last year. On its quarterly earnings call, executives said the growth would be driven largely by investments in AI infrastructure, particularly data centers and the processors inside them.  

In a November blog post, Meta’s leadership said it plans to invest $600B in U.S. data centers by 2028. Even at the upper end of its 2026 spending forecast, reaching that target would require capital expenditures to rise further in subsequent years.

The capex surge also follows Zuckerberg’s January unveiling of Meta Compute, a “top level initiative” to coordinate and finance data center build-out and power acquisition. 

“Meta is planning to build tens of gigawatts this decade, and hundreds of gigawatts or more over time,” Zuckerberg wrote on Meta’s Threads platform.  “How we engineer, invest, and partner to build this infrastructure will become a strategic advantage.”

Meta’s leaders said the company remains constrained by a lack of computing power, a shortfall they argued is limiting product development and driving its surge in capital spending.

The additional capacity is already contributing to revenue growth by improving user engagement and advertising performance across Meta’s social media products, Meta executives said. Much of the company’s AI investment is focused on strengthening recommendation systems for products like Instagram, enabling Meta to better infer users’ interests and match them with content and advertising.

But Meta’s AI ambitions go beyond improving how content and advertising are delivered on social media. Zuckerberg has touted a vision centered around “personal superintelligence,” agentic AI that mimics or exceeds human abilities. 

The CEO has remained tight-lipped about the products that will be enabled by these new models. But he said a wave of new AI products and tools will be unveiled in the months ahead. With capacity constraints expected to dissipate by the end of the year, the coming months could deliver a verdict on whether the fruits of the firm’s infrastructure surge justify the spending. 

“We are now seeing a major AI acceleration. I expect 2026 to be a year where this wave accelerates even further on several fronts,” Zuckerberg said. “We've been rebuilding our AI effort, and we're six months into that and I'm happy with how it's going, but we are going to be rolling out our initial set of models and products and businesses around that over the coming months.”

Calculating exactly how much Musk’s empire plans to spend on data centers and other AI infrastructure is more difficult.

On paper, xAI, Tesla and SpaceX are separate businesses in different industries. In practice, the lines between them have long been blurred — the firms share resources and often act in concert. But that coordination has grown more formal.

On Thursday, Bloomberg reported SpaceX may merge with Tesla or xAI. That same day, Musk announced that Tesla is investing $2B in xAI for a funding round aimed at building new data centers. The investment follows a separate $2B infusion from SpaceX into the AI startup late last year.

This sudden consolidation and spending surge coincides with a broad reorientation of all three companies toward AI and the infrastructure to support it. Rather than a carmaker, a space logistics firm and a generative AI developer, Musk is reshaping his tech empire as a collaborative ecosystem focused on AI computing and automation. 

The entity most directly ramping up capital deployment for data centers is xAI. In less than three years, the Grok developer has gone from a fringe player in the data center space to the first company to deliver a single campus with more than a gigawatt of power. 

In January, xAI disclosed a $20B Series E funding round largely earmarked for the construction of new data centers. The company then confirmed plans to allocate that entire $20B to a data center build-out in Southaven, Mississippi. This new 2 GW campus is only a few miles from xAI’s Colossus data centers in Memphis.

Musk has also begun repositioning SpaceX as a digital infrastructure company beyond its Starlink satellite internet business. Since December, he has increasingly framed the company’s future around putting data centers in space to support AI computing. 

On Friday, SpaceX filed a request with the Federal Communications Commission to deploy a constellation of up to 1 million solar-powered satellites designed to function as AI data centers. Musk has privately described SpaceX’s planned trillion-dollar initial public offering as a potential vehicle to finance the space-based computing push, Axios reported.

“It's a no-brainer building solar-power data centers in space. ... The lowest-cost place to put AI will be space, and that will be true within two years, three at the latest,” Musk said last month.

Tesla is also being reoriented toward Musk’s AI vision.

Musk said on Tesla’s earnings call that he no longer views the company as an automaker but rather a provider of AI-driven automation centered around self-driving cars and the company’s humanoid Optimus robot. Through this transition, Tesla is shifting resources to an automation market that AI leaders like Nvidia CEO Jensen Huang envision as a primary demand driver for AI computing. 

Tesla announced last week that its capital expenditures in 2026 will double to more than $20B, with a significant share directed toward expanding production of battery storage systems.

Battery storage is becoming a critical component of data center energy systems as operators reduce their reliance on the grid, and Tesla is increasingly steering its battery manufacturing capacity toward supplying those projects, including its own.

Musk also wants Tesla to begin producing its own silicon chips for use in its data centers and consumer products, outlining plans for a U.S.-based fabrication operation labeled TeraFab. Development of such manufacturing facilities typically requires tens of billions of dollars that Musk says weren’t included in Tesla’s capex projections. 

It’s a project Musk says is necessary to realize his AI ambitions. 

“We’ve got two choices: hit the chip wall or make a fab,” he said.