Anthropic's Revenue Run Rate Tops $65 Billion as IPO Talk Intensifies

Anthropic’s annualized revenue run rate topped $65 billion at the end of July, according to a source familiar with the figures reported by Reuters on August 17 — nearly double what rival OpenAI is tracking, and the latest datapoint in what has become the fastest revenue ramp in software history.

The number lands less than three months after the company disclosed that run-rate revenue crossed $47 billion at the time of its Series H close in late May. In other words, Anthropic added roughly $18 billion of annualized revenue in about two months. That is a pace of new ARR creation that no enterprise software company — not even Microsoft at its cloud peak — has ever sustained.

The quarter behind the number

The $65 billion figure is anchored by preliminary second-quarter revenue of more than $11.5 billion, first reported by The Information and confirmed by CNBC on August 15. That is up from $787 million in the same quarter a year ago — at least a 14-fold increase, and by some accounts closer to 15x. Annualized, a quarter above $11.5B puts the math squarely in line with the $65B+ run rate the company is now telling prospective investors about.

The trajectory across 2026 has been almost vertical:

• Roughly $9 billion run rate exiting 2025
• $14 billion by February 2026
• Around $19–20 billion by March
• $30 billion in April
• $47 billion at the Series H close in May
• $65 billion+ by end of July

By any conventional benchmark this is unprecedented. SaaS experts have pointed out that Anthropic is on track to out-earn every public software company except Microsoft by year-end — a company founded in 2021.

Why the market is watching

The revenue disclosure is not happening in a vacuum. It comes ahead of a widely reported initial public offering that Anthropic has been laying the groundwork for, having confidentially filed earlier this summer. The Information reports the company is forecasting $70 billion in annual revenue and $17 billion in cash flow by 2028 — projections that, if realized, would place it among the most valuable technology listings ever attempted.

The backdrop matters too. Anthropic closed a $65 billion Series H in May at a $965 billion post-money valuation, more than doubling its previous $380 billion mark, making it one of the largest private rounds on record. Alphabet, which invested $3 billion between 2023 and 2025 and later committed up to $40 billion more under a milestone-based structure, has seen its stake swell to roughly $124 billion according to Bloomberg — arguably the most lucrative venture bet in Google’s history. Amazon, meanwhile, has put in $8 billion with reports of up to $25 billion more.

What’s driving the growth

Three structural forces explain the ramp better than any single product cycle:

Enterprise coding agents. Claude’s dominance in agentic coding — through Claude Code, Claude for Work, and API usage embedded in tools like Cursor — has made Anthropic the default model vendor for the single hottest enterprise AI workload of 2026. Developers burn tokens at industrial scale, and coding agents run multi-hour autonomous sessions.

API-first distribution. Unlike consumer-first rivals, a large share of Anthropic’s revenue arrives through API calls where usage compounds quietly. With 300,000+ business customers reported by Sacra, the base is broad enough that growth no longer depends on any single logo.

Compute partnerships. The April agreement with Google and Broadcom for multiple gigawatts of next-generation TPU capacity gives Anthropic a supply runway that lets it say yes to demand — historically the binding constraint on AI lab revenue, not customer interest.

Caveats beneath the hype

Run-rate revenue is a projection, not a guarantee: it takes the latest period and multiplies it forward, assuming momentum holds. Gary Marcus and other skeptics have noted that the $47B figure from May was effectively Q2 (just over $11.5B) times four — treating a fast-growing quarter as if it were steady state. The same caveat applies to $65B. If Q3 growth merely flattens rather than accelerates, realized annual revenue will land well below the run-rate headline.

There are also cost questions. Serving billions in revenue requires enormous inference infrastructure, and Anthropic’s $17 billion cash-flow forecast for 2028 embeds aggressive assumptions about gross margins as context lengths and agent workloads grow. The bull case and the bear case both run through the same variable: how cheaply Claude can serve a token in 2027.

And competition is not standing still. OpenAI, Google, and a resurgent open-source ecosystem are all fighting for the same agent workloads, and price pressure at the frontier remains intense.

What it means

For the AI industry, Anthropic’s numbers are the strongest evidence yet that foundation-model companies can convert research leadership into durable, software-grade revenue at global scale — and do it fast. For investors, the pending IPO will be the first real public-market test of how durable frontier-lab economics look when subjected to quarterly scrutiny.

One thing is unambiguous: the gap between AI’s boosters and doubters used to be about whether the revenue exists at all. In August 2026, with Anthropic adding billions in annualized revenue per month, that debate has moved on to how long the slope can hold.