Anthropic told investors in June 2026 that its annualized revenue run rate had crossed $47 billion, a figure that more than doubles the roughly $20 billion pace OpenAI was reportedly generating earlier this year. The disclosure, part of a new funding round that valued Anthropic at $965 billion, sharpens the stakes for both companies as they prepare for what could become the two largest AI initial public offerings in history.
The number landed with force. Just months earlier, Anthropic had closed its Series G round at a $380 billion valuation, raising $30 billion in the process. Reaching $965 billion in the next round means the company’s private-market price roughly two-and-a-half-folded between fundraises, a trajectory that even by Silicon Valley standards is extraordinary.
Where the money is coming from
Claude, Anthropic’s flagship AI assistant, is the engine behind the growth. The model has moved well beyond developer sandboxes and into enterprise production environments. In February, Goldman Sachs announced a partnership with Anthropic to automate selected banking workflows using AI agents, giving the company a high-profile foothold on Wall Street. Deals like that one suggest a meaningful share of Anthropic’s revenue is coming from large institutions willing to embed AI into core operations, not just from API calls by startups experimenting with prototypes.
Anthropic’s deep relationship with Amazon has also been a distribution accelerator. Amazon Web Services offers Claude models through its Bedrock platform, giving Anthropic access to thousands of enterprise customers who already run workloads on AWS. Amazon has invested billions in Anthropic across multiple rounds, and the cloud channel likely accounts for a significant, though undisclosed, portion of the revenue run rate.
The OpenAI comparison
The rivalry at the center of this story hinges on a moving target. OpenAI disclosed an annualized revenue run rate of roughly $12.7 billion in late 2024. By early 2026, multiple reports pegged the company’s pace closer to $20 billion, driven by ChatGPT subscriptions, enterprise API contracts, and new product launches. Anthropic’s $47 billion figure, if it holds, would represent more than double that rate.
But precision matters here. Neither company has filed audited financials with the SEC. Run rates extrapolate a short window of sales, sometimes a single strong month, across a full year. A large upfront enterprise contract or a seasonal spike in API usage can inflate the snapshot. Until both companies file IPO prospectuses, the comparison is directionally useful but not airtight.
What is clear is that the gap closed, and then reversed, faster than most observers expected. A year ago, OpenAI held a comfortable revenue lead and a higher valuation. Anthropic has now overtaken it on both metrics in the private market, a shift that will force OpenAI to update its own narrative before pricing shares.
What the valuation actually means
At $965 billion, Anthropic’s private-market valuation places it alongside the world’s most valuable technology companies, ahead of where Meta and Amazon traded for years as public firms. But a private valuation set by a small group of investors in a funding round is not the same as a market capitalization determined by millions of daily trades. The number reflects what a handful of backers were willing to pay for a minority stake, often with protective terms like liquidation preferences that cushion downside risk.
That distinction will matter enormously at IPO. Public investors will demand audited revenue, operating margins, cash burn rates, and forward guidance. None of those figures are available today. AI model training and inference remain extraordinarily expensive, and there is no public evidence that Anthropic is profitable or close to breakeven. The $47 billion top line is impressive, but without knowing what it costs to generate, the valuation remains a bet on future margin expansion rather than a reflection of current earnings power.
Open questions that will shape the IPO
Several gaps in the public record deserve attention. First, customer concentration: if a handful of contracts, say Amazon’s cloud channel and a few large banks, account for the bulk of revenue, the growth story carries more risk than if thousands of mid-market companies are paying for Claude. Anthropic has not disclosed customer counts, average contract values, or churn rates.
Second, competitive pressure is intensifying from multiple directions. OpenAI continues to ship new models and expand ChatGPT’s capabilities. Google’s Gemini models are deeply integrated into Search, Workspace, and Cloud. Meta’s open-source Llama family has gained traction with developers who want to run models on their own infrastructure without per-token API fees. Anthropic’s ability to defend premium pricing against both proprietary and open-source alternatives will be a central question for IPO investors.
Third, the Goldman Sachs partnership, while confirmed, lacks independent performance data. Whether the bank has moved beyond a pilot phase, what specific tasks are being automated, and how Goldman measures return on investment are all unknown. A signed deal is a proof of concept, not proof of repeatable enterprise value at scale.
What investors should watch next
The practical signal to wait for is an S-1 filing. When either Anthropic or OpenAI submits its IPO prospectus to the SEC, the document will contain the first independently verified look at costs, margins, customer concentration, and contract durability. Everything before that moment, including the $47 billion run rate and the $965 billion valuation, is prologue.
That does not make the current numbers meaningless. Revenue run rates disclosed during capital raises carry legal weight; misrepresenting them to investors would expose the company to serious liability. And the speed of Anthropic’s ascent, from a research lab founded by ex-OpenAI executives in 2021 to a near-trillion-dollar private company five years later, is a genuine reflection of how fast the commercial market for large language models has matured.
But maturation is not the same as stabilization. The AI industry is still in a phase where demand is surging, pricing power is untested over full business cycles, and the cost structure of running frontier models at scale remains punishing. Anthropic’s $47 billion run rate is the strongest evidence yet that the company has built something customers will pay for. Whether they will keep paying, at these prices, in a market with more options every quarter, is the question that only time and audited books can answer.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
