Meta Fired 8,000 Workers. Now, Zuckerberg Signals 'Cloud War' With Amazon

For years, Meta stayed away from one of Big Tech's most profitable businesses -- cloud computing.

Amazon built Amazon Web Services (AWS). Microsoft expanded Azure. Google scaled Cloud. Even AI startups today rent computing power like electricity.

Meta did not. But that may be changing now.

Just days after cutting nearly 8,000 jobs globally, Mark Zuckerberg hinted that Meta could eventually enter the cloud computing business. The signal came during the company's annual shareholder meeting, where Zuckerberg admitted that competing with Amazon Web Services (AWS), Microsoft Azure, and Google Cloud is “definitely on the table”.

The statement may sound casual. It is not.

It reveals how Meta's aggressive AI strategy is slowly reshaping the company itself -- from a social media giant into an infrastructure heavyweight.

And the transformation is coming at an enormous cost.

First Came The Layoffs

Earlier this month, Meta laid off around 8,000 employees globally as part of a broader restructuring exercise. The company also cancelled nearly 6,000 open roles while moving another 7,000 employees into AI-focused teams.

The cuts were not random. Meta has been aggressively redirecting money toward artificial intelligence infrastructure. Data centres, AI chips, networking equipment, and computing clusters now sit at the heart of Zuckerberg's strategy.

The company has repeatedly said it wants to become an AI-first organisation. That transition is expensive. Very expensive.

Meta now expects to spend between $125 billion and $145 billion this year alone on AI-related capital expenditure. Earlier estimates were significantly lower.

The spending surge has already rattled investors. Meta shares fell sharply in April after Wall Street began questioning whether the company was spending too aggressively on AI infrastructure.

But Zuckerberg appears convinced that the gamble is necessary.

Why Compute Has Become The New Oil

Artificial intelligence models do not run on thin air.

Every chatbot, image generator, recommendation engine, or AI assistant depends on massive computing infrastructure operating behind the scenes. These systems require specialised chips, enormous data centres, cooling systems, and uninterrupted electricity.

This is where cloud computing enters the picture.

Companies like Amazon, Microsoft, and Google rent computing power to businesses worldwide. Instead of building expensive servers themselves, firms simply buy cloud access.

Today, AI companies are desperately hunting for more compute capacity. Anthropic CFO Krishna Rao recently described compute as “the lifeblood” of the AI business. That line captures the mood across the industry.

Demand is exploding. And Meta suddenly finds itself sitting on mountains of computing power.

The Question That Changed Everything

During Meta's shareholder meeting, Zuckerberg revealed something interesting. Companies are already approaching Meta asking whether they can buy compute from the firm.

“Almost every week,” Zuckerberg said, firms ask Meta whether it can provide API services or lease computing capacity. That matters because Meta currently does not operate a public cloud business.

Unlike AWS or Azure, Meta has historically built infrastructure only for itself -- to power Facebook, Instagram, WhatsApp, Threads, and now its AI ambitions.

But as Meta builds larger and larger AI clusters, the possibility of excess capacity grows. And that opens a new business opportunity.

Zuckerberg essentially admitted as much. “If we get to a point where we feel that we have overbuilt, then that is an option that we have,” he said.

In simpler terms: if Meta builds more AI infrastructure than it needs, it could start renting it out. Interstingly, that is exactly how AWS was born nearly two decades ago.

Meta Compute: The New Engine Inside Meta

At the beginning of this year, Meta quietly created a new division called “Meta Compute”. The unit is focused entirely on expanding the company's infrastructure footprint.

Zuckerberg has said Meta plans to build “tens of gigawatts” of capacity this decade and eventually scale toward “hundreds of gigawatts.”

The numbers are staggering. Meta already operates more than 30 data centres globally. It is also building giant AI facilities ranging between 1 and 5 gigawatts in size. One Louisiana project alone reportedly carries a price tag of roughly $27 billion.

The company is simultaneously building next-generation AI superclusters, including Prometheus and Hyperion. Hyperion is reportedly being designed at a scale comparable to Manhattan's footprint.

This is no longer normal tech expansion. It is industrial-scale infrastructure building.

Why Meta Wants A Backup Plan

There is another reason Zuckerberg keeps mentioning cloud computing: investors are nervous.

Meta still earns the overwhelming majority of its money from advertising. Yet it is now spending at levels usually associated with governments or semiconductor giants.

The company spent nearly $19 billion on capital expenditure in just the first quarter of 2026.

That pace has raised an uncomfortable question on Wall Street: what happens if AI revenues do not grow fast enough to justify these investments?

Cloud computing offers Meta a possible answer. If the company eventually commercialises excess infrastructure, its AI spending stops looking like a pure cost centre. Instead, it becomes a future revenue engine.

That narrative matters immensely for investors. AWS today generates more operating profit for Amazon than its retail business. Similarly, Microsoft's Azure has become central to the company's valuation story.

Meta clearly sees that opportunity. 

But Taking On AWS Won't Be Easy

Building data centres is one thing. Building a cloud business is another.

AWS, Microsoft Azure, and Google Cloud spent years creating enterprise sales teams, customer support systems, compliance frameworks, cybersecurity certifications, and developer ecosystems.

Meta currently lacks much of that infrastructure. The company is also deeply tied to some of the same players it may eventually compete against.

Meta reportedly has multi-billion-dollar agreements involving AWS, Google Cloud, CoreWeave, and Nebius. It has even signed cloud deals to support its own AI expansion.

That creates a strange dynamic. Meta is simultaneously becoming one of the world's biggest buyers of compute while exploring whether it could eventually become a seller too.

For now, Meta's cloud ambitions remain only a possibility. There is no launch timeline. No product announcement. No public roadmap.

But the direction is becoming clearer. First, Meta cut costs. Then it poured billions into AI. Now, Zuckerberg is openly talking about monetising the very infrastructure being built for that AI future.