Anthropic Raises $30 Billion, Valuation Exceeds $900 Billion
Summary: Anthropic (the developer behind the Claude large model) completed a new $30 billion funding round in May 2026, with a post-money valuation surpassing $900 billion, overtaking OpenAI as the world's most highly valued AI company. This round was jointly participated in by multiple sovereign wealth funds, tech giants, and strategic investors. The funds will be directed toward Claude model R&D, computing infrastructure expansion, and global market development. This article provides an in-depth analysis of Anthropic's fundraising logic, valuation support, technical moat, and the far-reaching impact of this financing round on the global AI industry landscape.
What Does $900 Billion Mean?
What does $900 billion mean?
On the global company market-cap leaderboard, Anthropic has surpassed Tesla (~$780 billion) and Berkshire Hathaway (~$850 billion), ranking behind only Apple, Microsoft, NVIDIA, Amazon, Google, and Meta — seventh in the world. And Anthropic is not even a publicly listed company — it has yet to achieve annual profitability.
Looking at the fundraising cadence, Anthropic's valuation curve is nearly vertical:
- Early 2023: ~$5 billion (Series C)
- Mid 2024: ~$18 billion (Series D, Amazon led with $4 billion)
- Early 2025: ~$60 billion (Series E)
- May 2026: ~$900 billion (Series F, $30 billion new funding)
From $5 billion to $900 billion in under three years — a 180× increase in valuation. This pace outstrips any technology company in history over a comparable period — including Cisco in 2000, Facebook in 2012, and Tesla in 2021.
But the numbers themselves are merely surface phenomena. The real question worth pursuing is: what justifies Anthropic's $900 billion valuation?
Who Is Anthropic? — More Than Just "The Claude Company"
Anthropic was founded in 2021 by former OpenAI VP of Research Dario Amodei and his sister Daniela Amodei. The founders' departure stemmed from disagreements with OpenAI over the "safety-first" approach — the Amodeis believed that under commercialization pressure, OpenAI was sacrificing the priority of AI safety research.
Consequently, Anthropic bore two distinctive labels from its inception:
- Safety first: Making AI alignment and controllability the core research objective, rather than commercialization priority
- Academic temperament: Research staff account for over 70% of the team, and the number of foundational research papers published ranks first among AI startups
This "academic-first" temperament once made capital markets cautious about Anthropic. But Claude 2's release in 2023 changed the game — Claude surpassed the contemporary GPT-3.5 across long-text understanding, logical reasoning, and safety, proving that "safety first" and "model capability" are not mutually exclusive.
After that, Anthropic entered the fast lane:
- 2024: Claude 3 series released, surpassing GPT-4 on multiple benchmarks
- 2025: Claude 4 released, with major improvements in reasoning and Agent capabilities
- 2026: Claude API's monthly active developers exceeded 2 million, enterprise customers surpassed 12,000
Composition and Strategic Intent of the $30 Billion Round
This $30 billion round is one of the largest single-round financings in AI history (second only to OpenAI's $35 billion round in 2025). Based on publicly disclosed information, the financing structure is as follows:
Investor Lineup
- Lead investors: Not fully disclosed, but known to include two sovereign wealth funds (reportedly Norway's sovereign fund and Singapore's GIC)
- Tech giants: Amazon continued to increase its stake, investing approximately $8 billion in this round; Google invested approximately $5 billion
- Strategic investors: NVIDIA, Microsoft (participated indirectly through a special-purpose entity)
- Financial investors: Multiple top-tier VCs followed on, including Lightspeed, Menlo Ventures, and others
One signal worth noting: Amazon and Google appear in the same funding round simultaneously, and both obtained board observer seats. This means Anthropic strategically maintains deep ties with both cloud giants — providing Anthropic with a unique "multi-cloud neutrality" advantage: neither dependent on a single cloud platform nor lacking in compute and distribution support from multiple giants.
Use of Funds
Anthropic disclosed three categories of fund allocation:
- Model R&D (~40%): Including development of the next-generation Claude 5, expansion of foundational research teams, and safety alignment research
- Computing infrastructure (~35%): Building proprietary compute clusters to reduce reliance on third-party clouds, and collaborating with NVIDIA on custom AI chips
- Global market expansion (~25%): Entering Asia-Pacific and European markets, building localized compliance systems, expanding enterprise sales channels
The computing infrastructure investment is the most noteworthy component. Anthropic launched a "proprietary data center" initiative in 2025, targeting the equivalent of 50,000 H100 GPUs of self-owned compute by 2027. This financing round will significantly accelerate that plan.
Valuation Support: Where Does the Confidence for $900 Billion Come From?
A $900 billion valuation corresponds to a price-to-sales ratio (P/S) of approximately 45× (based on Anthropic's estimated ~$20 billion revenue in 2025). This multiple far exceeds the 3–10× range for traditional SaaS companies and even surpasses NVIDIA's 20× P/E at the peak of the AI boom.
But capital markets apply a fundamentally different valuation logic to AI companies than to traditional technology companies. Four core logic chains support Anthropic's $900 billion valuation:
1. Model Capability as a Moat
Claude 4's performance across multiple authoritative benchmarks has established a "dual-hegemony" pattern alongside GPT-5. More importantly, Anthropic has built differentiated advantages in long-context understanding (2 million tokens) and safety and controllability — both critical in enterprise scenarios.
Enterprise customers choose Claude over GPT primarily because "Claude is less likely to produce harmful outputs" and "Claude's long-document analysis is more accurate." In high-risk industries like finance, healthcare, and law, these advantages translate directly into commercial value.
2. Depth of Enterprise Market Penetration
Anthropic's enterprise clients include JPMorgan Chase, Pfizer, Microsoft (yes, Microsoft is both OpenAI's largest shareholder and an Anthropic enterprise customer), NASA, and other heavyweight institutions. These clients' average annual contract value (ACV) exceeds $5 million, with renewal rates above 95%.
Enterprise market stickiness is extremely high — once Claude is deeply integrated into workflows, switching costs become prohibitive. This provides Anthropic with predictable long-term revenue streams.
3. Strategic Value of Compute
The core bottleneck for AI companies is not algorithms — it is compute. Through deep ties with Amazon, Google, and NVIDIA, Anthropic has secured priority compute supply rights. In the current global AI chip shortage, this "compute priority" is itself a massive competitive moat.
4. The Premium of the Safety Narrative
Anthropic is the only major AI company that positions "safety" as its core differentiator. In an increasingly regulated environment (EU AI Act, US executive orders), "safety compliance" is transitioning from a cost center to a revenue center — enterprise customers are willing to pay a premium for "safer AI."
The Ultimate Showdown with OpenAI
Anthropic's valuation surpassing OpenAI marks the AI industry's transition from "one superpower, multiple contenders" to "dual hegemony."
The competition between the two companies is essentially a contest between two philosophies:
| Dimension | OpenAI | Anthropic |
|---|---|---|
| Core positioning | Artificial General Intelligence (AGI) | Safe and controllable AI systems |
| Commercialization pace | Aggressive (GPT Store, Sora, Agent) | Measured (enterprise first, cautious on consumer) |
| Safety philosophy | Deploy and iterate | Verify before deploy |
| Fundraising strategy | Large-scale, high-visibility | Strategic binding, low-key expansion |
| Corporate culture | Product-driven, rapid iteration | Research-driven, long-termism |
In the short term, OpenAI still leads Anthropic in consumer influence and product richness. But Anthropic's advantages in the enterprise segment, research reputation, and safety compliance give it unique resilience in the long-term competition.
One interesting observation: within the AI researcher community (arXiv, Hacker News, academic conferences), Anthropic's research papers receive higher citation counts and more favorable reviews than OpenAI's. This accumulation of "academic reputation" is crucial in the long-term war for talent.
Impact on the Global AI Industry
Anthropic's financing round will have profound implications across three dimensions:
1. AI Bubble Discussions Will Be Reignited
A $900 billion valuation will inevitably trigger debates about whether AI is a bubble. But it is worth noting that Anthropic's revenue growth is equally astonishing — from approximately $4 billion in 2024 to roughly $20 billion in 2025, a year-over-year increase exceeding 400%. If this growth rate can be maintained in 2026, the valuation multiple will rapidly compress.
Historically, Amazon's P/S multiple exceeded 30× at the peak of the dot-com bubble in 1999, but subsequent revenue growth proved the valuation's合理性. Whether Anthropic will be the next Amazon or the next Cisco of 2000 depends on whether it can grow revenue beyond $100 billion within three years.
2. AI Infrastructure Arms Race Accelerates
Anthropic's massive investment in computing infrastructure will force OpenAI, Google, Meta, and other competitors to follow suit. The global scramble for AI chips, data centers, and energy will intensify further.
A direct consequence: AI compute costs are unlikely to drop significantly in the near term and may even continue rising due to surging demand. This poses a severe challenge for smaller AI companies that rely on third-party compute.
3. AI Safety Becomes Mainstream
Anthropic's success proves that "safety first" is not merely a moral stance but also a viable commercial strategy. This will encourage more AI startups to incorporate safety, alignment, and controllability into their core strategies, rather than treating them as "nice-to-have" add-ons.
The KaiheAiBox Perspective: Ecosystem Opportunities for the Agent Computer
Anthropic's fundraising bonanza is a strongly positive signal for the Agent Computer track.
The logic is straightforward: as underlying large models like Claude, GPT, and Gemini become increasingly capable and cheaper (Gemini 3.5's costs have already dropped 60%), the innovation space at the application layer expands dramatically. Large models are infrastructure; the Agent Computer is the "delivery vehicle" for large-model capabilities — packaging powerful AI into a product form that ordinary people can actually use.
KaiheAiBox's positioning is precisely this: it does not train large models but makes large-model capabilities truly work for people. Each generational improvement in underlying model intelligence expands what the Agent Computer can accomplish by an order of magnitude.
Anthropic's $900 billion valuation is essentially a bet on the belief that "AI capabilities will continue to grow exponentially." And the Agent Computer is the critical link that converts this belief into productivity.
Final Thoughts
$30 billion, $900 billion valuation — these numbers themselves are not what matter. What matters is the signal behind them: capital markets are betting on AI becoming the core of the next computing platform with unprecedented conviction.
Anthropic's story is not yet finished. Whether it can convert its valuation into sustained innovation capacity, whether it can maintain balance between commercialization and safety, whether it can carve a differentiated path in competition with OpenAI — these questions will be answered one by one over the next 3–5 years.
And for everyone paying attention to the AI industry, Anthropic's financing serves as a reminder: the pace of this transformation may be faster than we imagine.
KaiheAiBox | The Agent Computer for Everyone · AI Frontier Tracker