Anthropic has told investors it is on track for its first profitable quarter, projecting roughly $559 million in operating profit on $10.9 billion in revenue for the June quarter — figures disclosed privately to investors and reported by Bloomberg and The Wall Street Journal on May 20.
The $10.9 billion second-quarter projection more than doubles the roughly $4.8 billion Anthropic booked in Q1 2026. As recently as last summer, the company had told investors it did not expect full-year profitability before 2028. This quarter pulls that milestone forward by roughly two years.
What flips Anthropic into the black
The $559 million operating-profit figure includes the cost of training models but excludes stock-based compensation — the standard adjustment private AI labs use to present operating results. Anthropic also cautioned investors it does not necessarily expect to stay profitable in subsequent quarters: planned increases in compute and training spend can swing a single quarter back into the red. In other words, this is a profitable quarter, not yet a profitable company.
Even with those asterisks, the trajectory is unusual. Bloomberg reported the quarter-over-quarter growth rate outpaces the historical peaks of Zoom, Google, and Facebook at comparable stages.
Where the revenue is coming from
The surge is driven by enterprise demand for agentic tooling — Claude Code, the Claude API, and managed-agent deployments — rather than consumer subscriptions. Anthropic recently passed OpenAI in business-customer share by some third-party measures, and CFO Krishna Rao has pointed to a run-rate north of $30 billion. The revenue mix matters for buyers: growth concentrated in seat- and token-based enterprise contracts tends to expand rather than churn.
The IPO backdrop
The disclosure lands as Anthropic is reportedly in talks to raise at a $900 billion-plus valuation and weighing an IPO as early as October 2026. A demonstrable path to operating profit — even a lumpy one — strengthens the pitch to public-market investors who have grown wary of AI labs burning capital with no margin in sight.
What it means for builders
For teams standardizing on Claude, the signal is vendor durability: a supplier that can generate cash, at least intermittently, is less likely to gut free tiers or spike API pricing to plug a funding hole. The caveat cuts the other way too — Anthropic's own warning that compute spend will keep pressuring margins is a reminder that frontier-model pricing has a floor set by training and inference economics, not by competitive generosity. Capacity commitments and rate negotiations should assume the cost curve, not the headline profit, is the durable signal.
