Anthropic Hits $965 Billion Valuation, Overtaking OpenAI in the Race Toward a Trillion-Dollar AI Company

Anthropic, the AI company founded by former OpenAI executives Dario and Daniela Amodei, closed a $65 billion Series H funding round on May 28, 2026, at a post-money valuation of $965 billion [1][2]. The round was led by Altimeter Capital, Dragoneer Investment Group, Greenoaks, and Sequoia Capital [1]. For the first time, Anthropic's private-market valuation exceeds that of OpenAI, which closed its most recent round at $852 billion in March 2026 [4][5].

The milestone places Anthropic within striking distance of becoming the first private AI company valued at $1 trillion — a threshold that, just three years ago, few observers considered plausible for any startup.

Source: Bloomberg, CNBC

Data as of May 28, 2026CSV

The Fundraising Arms Race

Anthropic's ascent from a $4.1 billion valuation at its Series C in May 2023 to nearly $1 trillion in just three years represents one of the steepest valuation trajectories in corporate history [6].

Source: Bloomberg, TechCrunch

Data as of May 28, 2026CSV

The fundraising has accelerated sharply in 2026 alone. In February, Anthropic raised $30 billion in a Series G round at a $380 billion post-money valuation, led by Singapore sovereign wealth fund GIC and Coatue Management [7]. That round drew participation from a sprawling consortium including BlackRock, Blackstone, Fidelity, Goldman Sachs, JPMorgan Chase, Qatar Investment Authority, and Temasek [7]. Just three months later, the Series H more than doubled the company's valuation.

OpenAI's fundraising has been equally aggressive. Its latest committed capital round totaled $122 billion at the $852 billion valuation, anchored by Amazon ($50 billion), Nvidia ($30 billion), and SoftBank ($30 billion), with continued participation from Microsoft [8][9]. Over roughly the same period, OpenAI's valuation grew from $28 billion in April 2023 to $852 billion — a 30x increase [9].

The combined capital raised by these two companies in 2025–2026 alone exceeds $200 billion — a figure larger than the GDP of most countries.

Revenue: The Numbers Behind the Numbers

The valuation surge is not purely speculative. Anthropic's annualized revenue run rate crossed $45 billion in May 2026, up from $9 billion at the end of 2025 and just $87 million in January 2024 [10][11]. That growth rate — roughly 500x in under 30 months — has no precedent among technology companies. For comparison, Salesforce took approximately 20 years to reach $30 billion in annual revenue [11].

Source: VentureBeat, SaaStr

Data as of May 28, 2026CSV

OpenAI, meanwhile, is generating approximately $2 billion per month, or roughly $24 billion annualized, with 900 million weekly active users and over 50 million consumer subscribers [8].

At a $965 billion valuation on $45 billion in annualized revenue, Anthropic trades at roughly 21x revenue. At $852 billion on approximately $24 billion annualized, OpenAI trades at roughly 35x. Both multiples are high by historical standards — enterprise SaaS companies at scale typically trade at 8–15x forward revenue — but Anthropic's ratio is notably lower than OpenAI's, which may partly explain investor enthusiasm [12].

That said, annualized run rate is not the same as collected annual revenue. Run rate extrapolates the most recent month or quarter forward, and for companies growing this quickly, the figure can overstate sustainable revenue if growth decelerates.

Where the Money Comes From

Anthropic's revenue is heavily weighted toward enterprise and developer API access. Pay-per-token API consumption accounts for an estimated 70–75% of revenue, with consumer subscriptions (Claude Pro at $20/month, Claude Max at $100–200/month) making up 10–15% [13][14].

The enterprise traction is significant: over 300,000 business customers account for approximately 80% of revenue, with more than 100,000 running Claude through Amazon Bedrock [13]. The number of customers spending over $1 million annually doubled from roughly 500 to over 1,000 between February and April 2026 [13].

Claude Code, which became generally available in May 2025, has emerged as a major growth driver. The coding assistant hit $1 billion in annualized revenue by November 2025 and reached $2.5 billion in annualized revenue by February 2026, with enterprise customers including Netflix, Spotify, KPMG, and Salesforce [14].

Specific vertical breakdowns are limited, but enterprise contracts in healthcare, legal, and financial services are reportedly among the fastest-growing segments [13].

Who Owns Anthropic

The ownership structure of Anthropic involves some of the largest corporate and sovereign investors in the world, though governance control remains concentrated with the founders.

Google holds approximately 14% of Anthropic, having invested over $3 billion across multiple rounds, though it holds no voting rights, no board seats, and no board observer rights [15][16]. Amazon owns an estimated 7.8–19% of the company (estimates vary by source), having initially committed $8 billion — an investment now worth over $70 billion [17][18]. Like Google, Amazon has no governance control, no board seats, and no voting rights [17].

Both tech giants have benefited enormously from their stakes. A Fortune analysis found that unrealized gains on Anthropic holdings accounted for a significant share of Google's and Amazon's reported AI-related profits in recent quarters — raising questions about how much of Big Tech's "AI earnings" represent actual operating results versus paper gains on venture investments [18].

The remaining equity is distributed among dozens of institutional investors including Sequoia Capital, Fidelity, BlackRock, Blackstone, Goldman Sachs, and multiple sovereign wealth funds [7].

The Valuation Question: What Is $965 Billion Actually Worth?

Private-market valuations at this scale deserve scrutiny. Unlike public companies, whose share prices reflect continuous trading among millions of participants, private valuations are set by small, negotiated transactions among sophisticated insiders [12][19].

Several structural features of late-stage venture rounds can inflate headline numbers relative to the economic value available to common shareholders:

Liquidation preferences give preferred shareholders (investors) priority over common shareholders (employees, founders) in any exit scenario. When a company raises multiple rounds of preferred capital, these preferences stack. In many exit scenarios, preferred holders get paid first — common holders receive what remains, if anything [19].

Participation rights allow some preferred holders to receive their liquidation preference and then participate in remaining proceeds alongside common shareholders — sometimes called "double-dipping" [19].

Anti-dilution protections can reset investor share prices downward if a future round occurs at a lower valuation, shifting dilution onto common shareholders.

The practical implication: two companies with identical headline valuations can have radically different outcomes for employees depending on the preference stack [19]. Without access to Anthropic's term sheets, it is impossible to determine how much of the $965 billion figure flows through to common equity.

Generative AI platforms and infrastructure firms have been trading at 32–45x revenue multiples in private markets, a level that PitchBook's AIBQ scorecard has flagged as reflecting a "growing disconnect between valuations and fundamentals" [12].

Headcount and Efficiency

Anthropic operates with a remarkably lean team relative to its revenue and valuation. The company had approximately 2,300 employees as of December 2025, with estimates for early 2026 ranging from 2,900 to 5,000 depending on whether contractors and extended workforce are included [20][21].

At even the higher estimate, Anthropic generates roughly $9–14 million in revenue per employee — a figure that exceeds any company in the Forbes Global 2000 [20]. By comparison, OpenAI has approximately 3,700 employees and generates roughly $6.5 million per employee at its current run rate.

Google DeepMind, which operates as a division within Alphabet rather than a standalone revenue-generating entity, employs roughly 3,000 researchers and engineers. Meta AI similarly operates as a cost center within Meta, making direct comparisons difficult.

Anthropic's burn rate is not publicly disclosed. The company has raised over $40 billion in total funding across all rounds, and with revenue now running at $45 billion annualized, the gap between inflows and infrastructure spending is the key variable. AI model training runs at frontier scale cost hundreds of millions to billions of dollars each, and compute is the dominant expense line [22].

The Safety Brand: Substance or Strategy?

Anthropic was founded in 2021 when Dario Amodei, Daniela Amodei, and several colleagues departed OpenAI, partly over disagreements about safety practices and commercialization speed [23]. The company has since positioned itself as the "safety-first" AI lab, building its technical approach around Constitutional AI — a method for training models using a set of written principles rather than relying solely on human feedback [23].

This positioning has produced measurable technical differentiation. Anthropic's research on mechanistic interpretability — understanding what happens inside neural networks at a granular level — is widely regarded as among the most advanced in the field [23]. The company's Responsible Scaling Policy, now in its third version, establishes capability thresholds that are supposed to trigger additional safety evaluations before deployment [24].

Critics, however, question whether the safety framing is primarily a marketing and fundraising narrative. David Sacks, the Trump administration's AI czar, has called Anthropic's approach a "sophisticated regulatory capture strategy based on fear-mongering" [24]. Others have pointed out that Anthropic's RSP v3.0 replaced hard pause commitments with more flexible "graded public goals," and that the gap between published policies and actual deployment practices is not always clear at any of the major labs [24].

The counterargument from Anthropic's supporters: the company's Public Benefit Corporation structure and its Long-Term Benefit Trust (LTBT), which holds a special class of shares with escalating board-election rights, create governance mechanisms that no other AI company of comparable scale has adopted [25]. Over time, the LTBT is designed to give an independent body of technical and safety experts the ability to elect a majority of directors — a feature that could constrain profit-maximizing behavior in ways that traditional corporate governance does not [25].

Whether this structure represents genuine institutional innovation or an elegant mechanism that will prove toothless under commercial pressure remains an open question.

The IPO Horizon

Anthropic's fundraising activity is widely interpreted as groundwork for an initial public offering. Reuters has reported that the company hired law firm Wilson Sonsini to prepare for a possible IPO as early as late 2026, though a 2027 listing is considered more likely [25][26].

Several features of any Anthropic IPO would be unusual:

The Public Benefit Corporation designation means the company is legally required to balance shareholder returns with its stated public benefit — in this case, the responsible development of AI [25]. This is a similar but not identical structure to the one OpenAI has adopted as part of its for-profit transition.

The Long-Term Benefit Trust holding a special voting class would be among the most aggressive governance constraints ever brought to a public offering. Public investors have historically penalized dual-class share structures and trust-controlled governance, but Anthropic's management is expected to argue that the trust is what makes the safety brand credible [25].

Lock-up periods for insider shares would determine how quickly employees and early investors can sell after listing. Given the number of sovereign wealth funds and mega-cap institutional investors already in the cap table, secondary market dynamics could be volatile.

CoreWeave's March 2025 IPO offers a partial reference point. The AI infrastructure company priced at $40 per share — below its $47–55 target range — implying a $23 billion valuation roughly in line with its last private round [27]. The stock subsequently surged over 300% by mid-2025 before pulling back, demonstrating both the appetite for AI public offerings and the volatility that follows them [27].

What Comes Next

The competition between Anthropic and OpenAI is now playing out across multiple dimensions: revenue, valuation, model capability, enterprise adoption, and the race to public markets. Anthropic's overtaking of OpenAI on valuation, while symbolically significant, is a snapshot of a single negotiated transaction and could reverse with OpenAI's next fundraise.

The more durable question is whether either company — or both — can sustain revenue growth rates that justify valuations approaching $1 trillion while simultaneously spending tens of billions on compute infrastructure. The AI industry is in a period where revenue is growing faster than almost anyone predicted, but so are the capital requirements to stay competitive.

For investors considering Anthropic in a future public offering, the key variables are straightforward even if the answers are not: Can the revenue trajectory hold? Will the governance structure survive contact with public-market shareholders? And does a $965 billion valuation for a company founded five years ago represent a reasonable bet on AI's future — or the peak of a cycle that will look different in hindsight?