When Meta first outlined its capital expenditure plans last year, the message from leadership was unmistakable: artificial intelligence would sit at the center of the company’s future, and building the infrastructure to support it would require unprecedented investment. At the time, Meta’s Chief Financial Officer Susan Li made it clear that the company was prepared to spend aggressively to ensure it remained competitive in an increasingly crowded AI landscape.
“We expect that developing leading AI infrastructure will be a core advantage in developing the best AI models and product experiences,” Li told investors during an earnings call last summer, signaling that Meta was entering a new phase—one defined not only by software innovation, but by the scale and sophistication of the physical systems powering it.
Nearly a year later, Meta appears to be turning those projections into reality.
On Monday, Meta CEO Mark Zuckerberg announced the launch of Meta Compute, a sweeping new initiative aimed at dramatically expanding the company’s artificial intelligence infrastructure. The announcement marks one of the clearest indications yet that Meta is prepared to fundamentally reshape its energy usage, data center footprint, and long-term technology strategy in pursuit of AI leadership.
A Decade-Long Bet on AI Infrastructure
In a post shared on Threads, Zuckerberg outlined the scope of Meta Compute and the ambitions behind it. According to the CEO, Meta plans to build tens of gigawatts of computing capacity over the course of this decade, with the potential to scale to hundreds of gigawatts or more over time.
“Meta is planning to build tens of gigawatts this decade, and hundreds of gigawatts or more over time,” Zuckerberg wrote. “How we engineer, invest, and partner to build this infrastructure will become a strategic advantage.”
To put those figures into perspective, a single gigawatt represents one billion watts of electrical power—roughly enough to power hundreds of thousands of homes. By comparison, many large data centers today operate at capacities measured in the tens or hundreds of megawatts. Meta’s vision implies an infrastructure footprint on a scale more commonly associated with national power grids than individual corporations.
The push reflects the rapidly growing energy demands of generative AI systems, which require vast computing resources to train and operate large language models, image generators, and recommendation engines. As AI models grow in size and complexity, so too does the electricity required to power the servers running them.
Industry analysts have warned that AI could dramatically reshape electricity consumption patterns in the United States and beyond. Some estimates suggest that AI-driven power demand could surge from roughly 5 gigawatts today to as much as 50 gigawatts over the next decade, placing new strains on energy grids and prompting major investments in generation and transmission infrastructure.
From Social Media to AI Superpower
Meta’s aggressive infrastructure push underscores how far the company has evolved from its origins as a social networking platform. Once best known for Facebook and later Instagram and WhatsApp, Meta has increasingly repositioned itself as a technology company focused on AI, virtual reality, and large-scale computing.
Over the past several years, the company has invested heavily in AI research, open-source model development, and custom silicon. Its LLaMA family of large language models has become a cornerstone of its AI strategy, while internal tools powered by AI now influence everything from content ranking to advertising optimization.
However, as competition intensifies—particularly from Microsoft-backed OpenAI, Google, Amazon, and a growing ecosystem of AI startups—Meta faces mounting pressure to differentiate itself. Infrastructure, Zuckerberg suggests, may be one of the company’s most powerful competitive levers.
Leadership Team Behind Meta Compute
To execute on Meta Compute, Zuckerberg has assembled a leadership team that blends long-standing Meta veterans with newer hires who bring expertise from across the technology and policy worlds.
At the center of the initiative is Santosh Janardhan, Meta’s Head of Global Infrastructure. A company veteran who joined Meta in 2009, Janardhan has spent more than a decade helping build and scale the systems that support Meta’s global products.
According to Zuckerberg, Janardhan will oversee a wide range of responsibilities, including:
-
Technical architecture
-
Software stack development
-
Custom silicon programs
-
Developer productivity
-
The construction and operation of Meta’s global data center fleet and network
This broad mandate highlights how deeply infrastructure is now intertwined with Meta’s core business. Rather than being a background utility, infrastructure has become a central driver of innovation and performance.
Also playing a key role is Daniel Gross, who joined Meta last year after co-founding Safe Superintelligence, an AI-focused company he launched alongside former OpenAI chief scientist Ilya Sutskever. Gross brings a background that spans entrepreneurship, AI research, and long-term technology planning.
Zuckerberg said Gross will lead a new group within Meta responsible for:
-
Long-term capacity strategy
-
Supplier and vendor partnerships
-
Industry analysis
-
Infrastructure planning
-
Business modeling
The creation of this group reflects Meta’s recognition that AI infrastructure is not just a technical challenge, but also a strategic and economic one. Securing reliable suppliers, anticipating future demand, and managing costs at massive scale will be critical as the company expands.
The third key executive named is Dina Powell McCormick, a former government official who recently joined Meta as president and vice chairman. Her role will focus on working with governments and public-sector partners to help Meta “build, deploy, invest in, and finance” its infrastructure projects.
As data centers grow larger and more energy-intensive, collaboration with governments becomes increasingly important—particularly when it comes to permitting, energy sourcing, and regulatory compliance. Powell McCormick’s experience in public policy positions her to navigate these complex relationships.
Energy, Scale, and Strategic Advantage
At the heart of Meta Compute is a recognition that AI is fundamentally an infrastructure-driven business. While software breakthroughs often capture headlines, the ability to train and deploy advanced AI models increasingly depends on access to vast amounts of compute power, reliable energy, and efficient cooling systems.
Zuckerberg’s emphasis on energy scale reflects a broader industry shift. Major technology companies are racing to secure power generation capacity, explore renewable energy partnerships, and design data centers optimized for AI workloads.
Meta has already made significant investments in renewable energy over the past decade, often touting its commitment to sustainability. However, the scale implied by “tens of gigawatts” raises new questions about how such growth can be reconciled with environmental goals and grid limitations.
Industry observers note that AI infrastructure expansion could accelerate innovation in clean energy, storage, and grid optimization—but it could also intensify competition for power in regions already facing shortages.
A Competitive AI Arms Race
Meta’s announcement comes amid a broader race among technology giants to build AI-ready cloud environments. Capital expenditure projections released last year showed that many of Meta’s peers share similar ambitions, even if their strategies differ.
Microsoft, for example, has aggressively partnered with AI infrastructure providers while deepening its relationship with OpenAI. The company has invested heavily in expanding Azure data centers to support both internal AI development and customer demand.
Google’s parent company, Alphabet, has also made infrastructure a priority. In December, Alphabet announced the acquisition of data center firm Intersect, signaling its intent to further strengthen its computing backbone.
Amazon, meanwhile, continues to expand AWS with AI-specific chips and services, while investing in power-efficient data center designs.
Against this backdrop, Meta Compute can be seen as both a defensive and offensive move—defensive in ensuring Meta is not left behind, and offensive in positioning infrastructure as a differentiator rather than a commodity.
What Comes Next
While Zuckerberg’s announcement outlined the vision and leadership behind Meta Compute, many details remain unclear. Meta has not yet disclosed where new data centers will be built, how projects will be financed, or how the company plans to balance energy demand with sustainability commitments.
TechCrunch reached out to Meta for additional information about the initiative, but the company has not yet provided further comment.
What is clear, however, is that Meta is making one of the most ambitious infrastructure bets in the tech industry today. By committing to an energy footprint measured in gigawatts, the company is signaling that it views AI not as a short-term trend, but as a foundational technology that will shape its business for decades.
As AI systems become more deeply embedded in everyday life—from social media feeds and digital assistants to enterprise tools and creative platforms—the companies that control the underlying infrastructure may hold a decisive advantage.
With Meta Compute, Zuckerberg appears determined to ensure that Meta is one of them.
