Anthropic has achieved a $30 billion annual revenue run rate as of April 2026, up from approximately $9 billion at the end of 2025, making it the fastest-scaling B2B software company in history. This explosive growth, driven by enterprise adoption of its Claude AI model, has forced the company to commit $21 billion to secure nearly one million Google TPU (Tensor Processing Unit) chips from Broadcom, with delivery beginning in 2027. The infrastructure deal reveals a fundamental truth about frontier AI: the companies building these models must now operate at the scale of utilities, locking in compute capacity years in advance just to meet current demand .

Why Is Anthropic Growing So Fast?

Anthropic's growth trajectory has no parallel in software history. The company more than tripled its revenue in roughly three months, a pace that would be considered impossible in traditional enterprise software. The driver is clear: enterprise customers are embedding Claude into their core workflows at an unprecedented rate. More than 1,000 business customers are now spending over $1 million annually on Anthropic's services, a number that doubled in less than two months after the company closed its $30 billion Series G funding round in February 2026 .

A significant portion of this growth comes from Claude Code, Anthropic's coding assistant tool, which now generates over $2.5 billion in annual run-rate revenue. The tool's penetration into software development is measurable and accelerating: 4% of all public commits on GitHub are now authored by Claude Code, with projections reaching 20% or higher by year-end 2026. This isn't casual adoption; it's integration into the daily workflow of professional developers worldwide .

The enterprise focus distinguishes Anthropic from consumer-oriented AI competitors. While OpenAI has struggled with profitability despite $20 billion in annualized revenue, Anthropic is building a business model where enterprise customers pay directly for access. This shift from consumer subsidies to enterprise revenue creates a fundamentally different financial dynamic, though the capital intensity remains brutal .

How Does Anthropic's Infrastructure Strategy Differ From Competitors?

Rather than relying on a single cloud provider or chip vendor, Anthropic has deliberately built a multi-platform hardware strategy. Claude is trained and served across three distinct hardware ecosystems: Amazon's Trainium chips, Google's TPUs, and NVIDIA GPUs. This approach provides both resilience and negotiating leverage. If capacity becomes constrained on any single platform, workloads can shift. If one chipmaker faces supply disruption or pricing pressure, Anthropic is not exposed to the full shock .

The new Broadcom deal deepens Anthropic's Google relationship while maintaining its existing commitments to Amazon Web Services. Project Rainier, Anthropic's supercomputer cluster in Indiana, runs roughly 500,000 Amazon Trainium 2 chips and is expected to scale beyond one million chips by the end of 2025. The Google TPU capacity, set to come online in 2027, represents an additional layer of compute that will support future model iterations and inference workloads .

• Amazon Partnership: AWS remains Anthropic's primary cloud and training partner, with total Amazon investment reaching $8 billion and Project Rainier serving as the foundational supercomputer cluster.
• Google TPU Capacity: The new agreement secures approximately 3.5 gigawatts of next-generation TPU capacity starting in 2027, making it Anthropic's largest compute commitment to date.
• NVIDIA GPU Access: Claude continues to run on NVIDIA GPUs across Microsoft Azure, ensuring the company maintains optionality across all three major cloud platforms.
• Geographic Concentration: The majority of new compute capacity will be deployed in the United States, aligning with Anthropic's November 2025 commitment to invest $50 billion in domestic AI infrastructure.

What Does the $21 Billion Broadcom Deal Actually Secure?

The Broadcom agreement represents three distinct but interconnected commitments. First, Anthropic gains access to nearly one million Google TPU v7p units, the next-generation tensor processing chips optimized for AI workloads. Second, Broadcom has signed a separate long-term agreement with Google to design and supply future generations of custom TPU chips, ensuring a pipeline of hardware innovation. Third, Broadcom provides a supply assurance agreement for networking and other components needed for Google's next-generation AI data racks through 2031 .

Broadcom's role has evolved from component supplier to infrastructure integrator. The company now sells fully assembled "Ironwood Racks" directly to AI labs, providing a turnkey solution that reduces integration complexity and cost. This arrangement offers an alternative to NVIDIA's general-purpose GPUs, with some experts arguing that Google's TPUs are more efficient for specific AI algorithms. For Anthropic, the deal guarantees specialized, high-performance compute without relying on the volatile cloud market .

"This groundbreaking partnership with Google and Broadcom is a continuation of our disciplined approach to scaling infrastructure. We are building the capacity necessary to serve the exponential growth we have seen in our customer base while also enabling Claude to define the frontier of AI development," said Krishna Rao, Chief Financial Officer at Anthropic.

Krishna Rao, Chief Financial Officer at Anthropic

The financial commitment is substantial but strategically calculated. The $21 billion is a capital expenditure that will be amortized over the lifespan of the chips, not immediate cash burn. It represents a prepayment for capacity that will underpin revenue growth for years. Analysts at Mizuho estimated that Broadcom alone would record $21 billion in AI revenue from Anthropic in 2026, rising to $42 billion in 2027, illustrating the financial weight of the commitment .

Can Anthropic Actually Reach Profitability?

The path to profitability for frontier AI companies remains uncertain and heavily dependent on execution. Anthropic projects positive cash flow by 2027, a target that hinges on its enterprise focus and the successful scaling of its $30 billion run-rate revenue. However, the capital intensity of the industry is undeniable. The AI sector is expected to spend $690 billion in capital expenditure in 2026 alone, a figure that underscores the war chest required just to maintain competitive parity .

The contrast with OpenAI is instructive. OpenAI projects a $14 billion loss in 2026 despite $20 billion in annualized revenue, a gap driven by the fact that only 5.5% of its massive user base pays for access. The company is funding a capital-intensive race with a consumer model that burns cash. Anthropic's enterprise-first approach avoids this trap by generating revenue directly from customers who depend on Claude for business operations, not casual use .

The multi-cloud strategy provides a critical advantage in managing capital efficiency. By training and running models across AWS Trainium, Google TPUs, and NVIDIA GPUs, Anthropic can match workloads to the most efficient hardware for each task. This flexibility is a defensive moat against giants like Microsoft and NVIDIA, which control vast cloud and chip ecosystems. However, executing this strategy efficiently is paramount, and any misstep in hardware selection or capacity planning could derail the profitability timeline .

Steps to Understand Anthropic's Competitive Position in Enterprise AI

• Revenue Growth Metric: Track Anthropic's quarterly revenue run-rate and compare it to OpenAI's reported figures to assess which company is winning the enterprise market.
• Customer Concentration: Monitor the number of enterprise customers spending over $1 million annually; Anthropic doubled this figure to over 1,000 in less than two months, signaling accelerating adoption.
• GitHub Penetration: Watch Claude Code's share of public GitHub commits as a proxy for developer adoption; the tool currently accounts for 4% of commits with projections to reach 20% by year-end.
• Compute Capacity Announcements: Follow infrastructure deals and capacity commitments from all frontier AI labs; these reveal which companies have secured the physical resources to scale and which are facing bottlenecks.
• Profitability Timeline: Assess whether Anthropic achieves positive cash flow in 2027 as projected; this will validate whether the enterprise-first model can generate sustainable returns.

Anthropic's $30 billion revenue run rate and $21 billion infrastructure commitment signal a company operating at a scale and speed that redefines what is possible for a private B2B software business. The company is not just selling AI models; it is becoming the essential operating system for enterprise AI and software development. Its growth trajectory is a direct function of that ambition, and its ability to secure compute capacity years in advance is the physical manifestation of a business model that has achieved escape velocity .