Anthropic Surpasses OpenAI in Valuation, Hits $965B

The shift in valuation is more than a numbers game; it is a signal that the market is rewarding a different approach to AI scaling. While OpenAI built a centralized empire, Anthropic has spent the last few years positioning itself as the industry’s essential hedge. By diversifying its capital sources and compute partnerships, the company has effectively insulated itself from the volatility of any single cloud provider or strategic partner.

This latest round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia, with each lead investor contributing more than $2 billion, according to 凤凰网财经. The investor list reads like a directory of global institutional power, including Capital Group, D1 Capital, GIC, BlackRock, Blackstone, and Temasek. Even the hardware layer is now deeply integrated into the cap table, with memory giants Micron, Samsung, and SK Hynix providing both capital and strategic alignment.

The inclusion of these chipmakers is a calculated move to secure the physical foundations of the AI race. With the industry facing chronic shortages of high-bandwidth memory (HBM) and logic chips, these investments create a dual bond of equity and supply, ensuring Claude’s infrastructure can scale without being throttled by hardware bottlenecks.

The Swiss Bank Model: Multi-Cloud and Hardware Alliances

The strategic divergence between the two giants is stark. OpenAI operates on an “empire model,” tightly bound to Microsoft and the massive Stargate project. In contrast, 澎湃新闻 describes Anthropic’s approach as a “Swiss Bank model,” where the company avoids exclusive bindings to ensure it remains the safest bet for a variety of partners.

This strategy has allowed Anthropic to play the cloud giants against one another. Claude is now the first frontier model available across all three major cloud platforms: Amazon Web Services (AWS), Google Cloud, and Microsoft Azure.

• Amazon: A new agreement providing up to 5 gigawatts (GW) of capacity.
• Google and Broadcom: A deal for 5GW of next-generation TPU capacity.
• SpaceX: Access to GPU capacity via the Colossus 1 and Colossus 2 compute clusters.

The SpaceX deal is particularly telling. In a reversal of previous hostilities, Elon Musk has rented the Colossus 1 cluster—a beast containing over 220,000 NVIDIA GPUs—to Anthropic. For Musk, it is a pragmatic business move to monetize idle assets; for Anthropic, it is a massive injection of raw power that accelerates its training capabilities just as it prepares for its next major technical leap.

This multi-platform layout does more than just provide chips. It offers enterprise customers superior system stability and performance by allowing computational tasks to be distributed across the most compatible hardware, whether that be AWS Trainium, Google TPUs, or NVIDIA GPUs.

Revenue Acceleration and the Path to Profitability

Valuations in the AI sector have long been driven by hype and model benchmarks, but Anthropic is shifting the conversation toward “blood-making” capability—actual revenue generation. The company is no longer just a research lab; it is a commercial engine.

Financial projections indicate a vertical growth trajectory. According to 新浪财经, Anthropic expects its second-quarter revenue to hit $10.9 billion, more than doubling its previous quarterly performance. More impressively, the company has informed investors that its annual recurring revenue (ARR) run rate is expected to exceed $50 billion by the end of June 2026.

To put that growth in perspective, the company’s run rate was only $4 billion in July 2025. This trajectory suggests that Anthropic has successfully transitioned from a challenger to a dominant market force, capturing enterprise spend that was previously reserved for ChatGPT.

The company is also eyeing its first profitable quarter, a milestone that would separate it from many of its peers who are burning billions in a race for AGI. This financial discipline makes the company a highly attractive candidate for the public markets, with both Anthropic and OpenAI expected to potentially launch IPOs as early as this autumn.

The ASI Duel: Stargate vs. Colossus

As the industry moves toward Artificial Super Intelligence (ASI), the competition has evolved into a war of attrition fought with electricity and silicon. OpenAI is doubling down on its Stargate project, a $500 billion investment commitment aimed at creating a compute capacity exceeding 10GW—roughly the output of five Hoover Dams, as reported by 第一财经.

OpenAI’s strategy is one of absolute scale. It has secured $122 billion in recent financing, with massive bets from Amazon ($50 billion), NVIDIA ($30 billion), and SoftBank ($30 billion), to push its valuation to $852 billion. It is building a sovereign-level compute infrastructure designed to dominate the next era of intelligence.

Anthropic is fighting this with agility and technical refinement. Along with its funding, the company released Claude 4.8 Opus, a flagship model designed to bridge the gap to IPO. This version focuses on architectural leaps in multi-agent coordination and engineering implementation, targeting the high-end enterprise market where reliability and “agentic” workflows are more valuable than raw chat capabilities.

The stakes are no longer just about who has the best chatbot. This is a battle over the infrastructure of the future economy. OpenAI is building a centralized utility; Anthropic is building a distributed network of alliances. If OpenAI’s Stargate succeeds, it creates a monopoly on ASI. If Anthropic’s “Swiss Bank” model wins, it creates a fragmented but resilient ecosystem where the most flexible player captures the most value.

With the valuation lead now in Anthropic’s hands and a revenue run rate climbing toward $50 billion, the “underdog” has effectively ended the era of OpenAI’s uncontested dominance. The coming months—and the potential autumn IPOs—will determine which of these two architectural philosophies the public markets truly value.