Elon Musk is transforming xAI and SpaceX into cloud infrastructure providers, betting that renting out AI computing power to rivals can help fund his massive AI ambitions. Rather than keeping all that computing capacity for his own models, Musk signed a $40+ billion contract with Anthropic, his arch rival, to let the company use Grok and Colossus data center infrastructure to serve their models. This move mirrors a broader industry shift where companies with massive AI infrastructure investments are now monetizing excess capacity.

The strategy reflects a fundamental reality of the AI boom: building frontier AI models is extraordinarily expensive, and companies are finding creative ways to offset those costs. Meta is pursuing a similar path, explicitly launching an enterprise cloud computing business on top of its multi-hundred-billion-dollar AI data center buildout. Mark Zuckerberg confirmed that a Meta cloud computing business is "definitely on the table," according to recent reporting.

Mark Zuckerberg

Why Are Tech Companies Suddenly Becoming Cloud Providers?

The answer lies in the sheer scale of AI infrastructure investment. Meta alone is spending over $145 billion this year on AI data center infrastructure. While much of that capacity serves Meta's 3.5 billion-plus users across Facebook, Instagram, and WhatsApp, and powers its AI-driven advertising, there is significant excess capacity sitting idle. Rather than let that computing power go unused, companies are now selling access to enterprise customers, much like Amazon Web Services (AWS) does with general cloud computing.

This represents a fundamental shift in how AI companies think about their infrastructure. Historically, companies built data centers exclusively for their own products. Now, the economics of AI have become so demanding that sharing infrastructure with competitors makes financial sense. Musk's deal with Anthropic is particularly striking because the two companies are direct rivals in the race to build the most capable AI models.

How Are Companies Structuring Their AI Cloud Strategies?

The industry is developing what one analyst calls a "Goldilocks" approach to AI, with three distinct strategies emerging across the technology landscape:

• Frontier Model Companies: Anthropic and OpenAI are focused on rapid iteration of their flagship large language models (LLMs), which are AI systems trained on vast amounts of text to understand and generate human language. Anthropic released Opus 4.8 just 41 days after Opus 4.7, with a more advanced Mythos-class model expected within weeks. OpenAI recently shipped GPT 5.5 ahead of its planned September initial public offering.
• Middle-Tier Infrastructure Players: Meta and SpaceX/xAI are leveraging their multi-hundred-billion-dollar data center buildouts into AWS-style enterprise plays. These companies are rolling out their own AI models, such as Meta's Muse Spark and Grok 5 from xAI, while simultaneously offering computing resources to other companies. This dual approach helps defray the enormous costs of both model development and infrastructure.
• Edge and Local AI Specialists: Apple is doubling down on running AI models directly on devices rather than in the cloud, leveraging its unparalleled global supply chain across 2 billion-plus devices. This strategy prioritizes privacy and low latency by keeping AI processing local to the user's phone, tablet, or computer.

These three approaches are not competing so much as coexisting, each optimized for different market segments and use cases. The frontier model companies focus on raw capability and speed of innovation. The middle-tier players balance their own model development with revenue generation from enterprise customers. Apple focuses on seamless integration across its ecosystem and on-device privacy.

What Does This Mean for Competition in Cloud Computing?

The emergence of Meta and SpaceX/xAI as cloud infrastructure providers expands the competitive landscape beyond the traditional "big three" of Amazon AWS, Microsoft Azure, and Google Cloud. This shift is significant because AI compute is becoming one of the most essential components of the technology stack as artificial intelligence scales globally throughout the decade.

Margins and growth in this segment remain high for now and should maintain momentum in the near term. However, the entry of well-capitalized competitors like Meta and SpaceX could eventually pressure pricing and force consolidation. The fact that Musk is willing to let Anthropic use his infrastructure demonstrates how critical it has become to monetize these massive investments. Meta's explicit pivot toward enterprise cloud services signals that the company sees this as a core business opportunity, not merely a cost-recovery mechanism.

These trends are still in their early phases and will likely play out through 2030 as the industry matures. For now, the "Goldilocks" strategies across frontier models, middle-tier cloud services, and edge AI represent a rational division of labor in an industry grappling with unprecedented infrastructure costs and competitive pressure.