KEY POINTS
- Anthropic's annualized revenue reached $30 billion, surpassing OpenAI's $25 billion for the first time, at an $800 billion valuation.
- Anthropic is targeting a U.S. listing as early as October 2026 and a raise of more than $60 billion, which would make it the largest AI IPO on record.
- OpenAI is still targeting a Q4 2026 NASDAQ debut at roughly $852 billion, but secondary-market demand is quietly tilting toward Anthropic.
Anthropic's annualized revenue hit $30 billion on April 7, surpassing OpenAI's reported $25 billion for the first time and pushing its implied private valuation to roughly $800 billion, according to secondary-market pricing reported by TradingKey and Fortune. The company is targeting a U.S. listing as early as October 2026 and is in early talks to raise more than $60 billion — a figure that, if delivered, would rank as the largest technology IPO ever printed. OpenAI, meanwhile, closed a $120 billion funding round in March at an $850 billion post-money valuation and is still aiming for a Q4 2026 NASDAQ debut, according to a CNBC report on the competitive public-markets setup. For the first time in three years, the narrative flipped.
The Enterprise Share Shift
The revenue crossover was not an accident. Anthropic's enterprise market share rose to 30.6% from 24.4% in the prior survey, according to data summarized in an April TechCrunch piece, and the company now posts a 70% win rate against OpenAI on new enterprise buyers. That reflects a shift in how Fortune 500 CIOs think about AI risk. Two years of regulatory pressure around model reliability, data handling and agentic behavior have pushed procurement committees toward the vendor that positions safety as a product, not a slogan. Claude's traction in regulated verticals — banking, healthcare, legal services — now translates directly into multi-year ACV commitments that are easier to underwrite than OpenAI's consumer-heavy mix.
Margin path matters too. Anthropic is projecting GAAP profitability by 2028. OpenAI, by contrast, continues to carry a longer breakeven runway driven by compute commitments, consumer acquisition costs and a sprawling product footprint. For a private-to-public transition at a nearly trillion-dollar valuation, the profitability slope is the number the IPO book builds around.
The Secondary Market Is Already Voting
Secondary markets tend to front-run narrative shifts. Demand for Anthropic shares in private marketplaces has turned what investors describe as "nearly insatiable," with secondary trades clearing at or above the $800 billion implied mark. OpenAI shares, by contrast, have been trading at a discount to the $850 billion primary round in some venues — a signal that some inside investors are getting nervous about concentration and the cost structure of the next capex wave.
That wave is not slowing. U.S. hyperscaler AI capex is on pace to approach $700 billion in 2026, and Microsoft, Google, Amazon and Meta each raised their infrastructure budgets on their most recent calls. Anthropic benefits doubly: it earns API revenue on Claude usage across Amazon Bedrock and Google Cloud Vertex, and it owns its own training and inference stack through partnerships with AWS Trainium and Google TPU. OpenAI's compute remains more Microsoft-dependent, and the recent capacity tension between Microsoft and OpenAI around the evolving Stargate buildout has added a governance wrinkle that public investors will pick apart in an S-1.
The Public Opinion Overhang
Both companies share one meaningful problem heading into their offerings. Public sentiment on AI and data centers has turned sharply negative in the most recent polling cited by CNBC, with concerns around energy consumption, job displacement and content provenance driving state-level legislation. More than 600 AI bills were introduced in 2026 legislative sessions, and California, New York and Texas are all moving on companion-chatbot disclosure and synthetic-content labeling. A listing at $800 billion or more sits atop a regulatory base that is visibly shifting.
The Trump administration's March National Policy Framework for AI leaned toward preemption of state-level laws it considered unduly burdensome, which partially offsets the risk. But enforcement mechanics remain unresolved, and the Brazilian antitrust probe into Microsoft's Jumpstart program shows that international exposure is a live book-building issue for anyone going public in the current environment. Europe's AI Act implementation enters a harder enforcement phase in the second half of 2026, and both companies will need to disclose jurisdiction-specific compliance posture in their prospectuses. Public investors who have not sat through a regulated-tech IPO since the 2021 cohort will need to price that complexity in.
For traders, the takeaway is specific. The publicly listed AI exposures — Microsoft, Nvidia, Alphabet, AMD, Arm, TSMC — will continue to act as the liquid proxy for whichever of these two companies establishes market leadership. Anthropic pulling ahead in enterprise share directly benefits Amazon and Google, Anthropic's cloud distribution partners. OpenAI still benefits Microsoft disproportionately, though that leverage is softening with every incremental Claude deployment. The margin flow-through matters: hyperscaler AI revenue at 60%-plus gross margins is now the single biggest driver of consensus earnings revisions for Microsoft and Alphabet into late 2026.
Watch the IPO filing cadence into the fall. An Anthropic S-1 in October would likely come with a greenshoe-heavy deal and a float well under 10%, creating the kind of post-listing supply tightness that marked the 2019 CrowdStrike and 2020 Snowflake debuts. The key date is the NYSE/NASDAQ confidential filing window, which would need to open by late July to support an October print. Between now and then, the cleanest tell is Anthropic's next enterprise win-rate survey and whether OpenAI's Q3 revenue run-rate closes or widens the gap. Traders looking for optionality should also watch the behavior of Microsoft and Alphabet shares on any confirmed Anthropic S-1 headline — a wide divergence between the two would tell you the market has finally decided which hyperscaler wins the AI distribution fight, and that call has been pending for a full year.
