Fifteen months ago, Anthropic was generating roughly $1 billion a year selling access to its Claude AI models. By early April 2026, that figure had exploded to a $30 billion annualized revenue run rate, according to Bloomberg. If the pace holds, the San Francisco startup founded by former OpenAI researchers has surpassed its older, better-known rival on top-line sales and pulled off what may be the fastest commercial scaling in the history of enterprise software.
For context, OpenAI was last publicly reported to be running at roughly $11.6 billion in annualized revenue in late 2025. Salesforce, the company that defined cloud software, needed more than 20 years to cross the $30 billion mark. Anthropic, led by CEO Dario Amodei, appears to have reached it in under four.
The deals driving the surge
Three moves in quick succession fueled the acceleration.
First, Anthropic closed a $30 billion Series G in February 2026, valuing the company at $380 billion post-money, as reported by Reuters. That round, stacked on top of roughly $8 billion Amazon had already poured into the company by late 2025, gave Anthropic a war chest few startups have ever held.
Second, Broadcom finalized a deal to ship Google TPU chips directly to Anthropic, Bloomberg confirmed in the same April reporting. The arrangement gives Anthropic dedicated silicon outside its existing cloud partnerships with Amazon Web Services and Google Cloud. Instead of renting compute by the hour, the company is building its own inference capacity, a shift that could improve margins as usage scales.
Third, Anthropic committed to investing $50 billion in domestic AI infrastructure, a pledge listed on the federal government’s AI action plan. The commitment covers data centers and chip-related construction on U.S. soil, tying Anthropic’s expansion to Washington’s push for onshore AI supply chains. No binding timeline has been disclosed, and corporate pledges of this scale often stretch over a decade, but the listing signals that policymakers view Anthropic as a strategic asset, not just a commercial one.
Wall Street is already buying in
Revenue at this scale does not come from chatbot subscriptions alone. Anthropic sells API access to its Claude family of large language models, enterprise licenses, and increasingly, AI agents that can execute multi-step tasks inside corporate workflows.
The most visible enterprise customer so far is Goldman Sachs. In February 2026, Reuters reported that the bank had partnered with Anthropic to deploy AI agents for automating internal banking tasks. Financial institutions are notoriously slow adopters of new technology; compliance reviews, vendor risk assessments, and regulatory scrutiny add months to procurement cycles. That Goldman moved forward suggests Anthropic’s models cleared a high bar for reliability and security in one of the most regulated industries on earth.
One bank deal, even a marquee one, does not explain a $30 billion run rate on its own. The bulk of that figure likely comes from a broader base of enterprise and cloud customers whose identities remain confidential, plus usage-based API revenue that can swing quarter to quarter. But the Goldman partnership matters as a signal: if the largest names in finance are integrating Anthropic’s agents into live systems, the pipeline of similar deals across banking, insurance, and asset management could be deep.
What the numbers do not tell you
A run rate is a snapshot, not a guarantee. It extrapolates a recent stretch of revenue, sometimes a single strong month, into an annual figure. If Anthropic landed a cluster of large one-time contracts or benefited from a seasonal surge in enterprise procurement, the $30 billion pace could overstate its steady-state business. Anthropic is private and has no obligation to publish audited financials, so there is no way for outsiders to verify quarterly trends, margins, or customer concentration.
Profitability is an even bigger unknown. Training and running frontier AI models is extraordinarily expensive. The Broadcom chip deal secures supply, but its pricing and volume terms have not been disclosed. If Anthropic is paying a premium to lock in scarce silicon, the impressive top line could be paired with thin or negative margins. The $50 billion infrastructure pledge, meanwhile, represents future capital expenditure that will weigh on cash flow for years.
Customer concentration is another blind spot. If a handful of hyperscale cloud partners, banks, or government contracts account for most of the revenue, Anthropic’s business is more fragile than the headline number suggests. A diversified base of thousands of smaller customers would be far more durable, but the reporting available does not reveal which scenario is closer to reality.
The comparison to OpenAI also carries a caveat. OpenAI’s most recent public revenue figures date to late 2025. The company may have grown substantially since then, and without simultaneous disclosure from both firms measured over the same window, any ranking is provisional.
What to watch through the rest of 2026
The next few quarters will determine whether Anthropic’s run rate reflects a durable business or a peak moment. Several things will clarify the picture:
- Quarterly revenue consistency. If leaks or investor updates show the $30 billion pace holding through mid-2026, the figure moves from impressive to structural.
- Margin signals. Any disclosure around cost of revenue, cloud spending, or chip procurement terms will reveal whether growth is translating into a viable business model or simply buying market share.
- Enterprise expansion beyond finance. Goldman Sachs opened the door. Watch for announcements in healthcare, legal, and government, sectors where compliance requirements are similarly high and where adoption would validate Anthropic’s enterprise playbook.
- OpenAI’s response. OpenAI is not standing still. Its own enterprise push, consumer products, and rumored fundraising will shape whether Anthropic’s lead is a gap or a momentary crossing.
- IPO or secondary-market pricing. At a $380 billion valuation, Anthropic is priced like a public mega-cap. Any move toward a listing would force the transparency that private markets currently do not require.
What is already clear is that Anthropic has moved from research lab to commercial juggernaut faster than almost anyone predicted. The verified milestones, a $30 billion run rate, a $380 billion valuation, a direct chip supply line, a Wall Street flagship customer, and a $50 billion infrastructure commitment, are each significant on their own. Together, they describe a company operating at a scale and speed that is reshaping the competitive landscape of AI in real time.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
