Anthropic Heads for First Profitable Quarter on $10.9B in Q2 Revenue

Anthropic is on pace to post its first operating profit ever in the June quarter, on revenue projected at $10.9 billion, according to figures the company shared with investors during its ongoing funding round. The numbers, reported by CNBC and confirmed by additional sources, arrive roughly two years ahead of Anthropic's own internal projections for breakeven.

The disclosure changes the financial story around the company. Until this week, Anthropic was widely understood as a high-burn lab spending aggressively on compute and pretraining. The Q2 figures suggest the spend is now being matched, and then some, by the revenue side of the business.

The numbers

Anthropic told investors that Q2 revenue will more than double from Q1's $4.8 billion to $10.9 billion. The company projects $559 million in operating profit on that revenue for the quarter. The operating profit figure includes model training costs but excludes stock-based compensation, a standard caveat for late-stage tech companies.

The quarter-over-quarter growth rate — roughly 127 percent — exceeds the historical quarterly peaks of Zoom, Google, and Facebook during their respective hypergrowth phases. That comparison is not a Anthropic talking point; it's how some investors close to the round are framing the magnitude.

Why this matters: profitability changes Anthropic's negotiating position with everyone. Customers stop worrying about whether the company will exist in three years. Hyperscalers stop treating Anthropic as a charity case for cloud credits. Investors stop pricing in burn risk and start pricing in earnings multiples.

Why earlier than expected

Last summer, Anthropic told investors it did not expect full-year profitability until at least 2028. Two things pulled the timeline forward.

The first is enterprise revenue. Claude's business adoption surged through Q1 and Q2 as companies moved from pilots into production deployments. Recent reporting put Anthropic ahead of OpenAI in business AI market share, which translates to more recurring revenue at higher gross margins than consumer subscriptions.

The second is API pricing power. With Claude Opus 4.7 launched as the flagship coding and vision model and Claude Sonnet 4.6 used by developers at scale, Anthropic has been able to hold pricing on premium tiers while consumer-facing competitors race down on cost. Enterprise customers, especially in regulated industries, tend not to shop on price for a model they've already integrated into production workflows.

The compute deals do not break the math

Anthropic signed two of the largest compute commitments in industry history this month. The company will pay roughly $1.25 billion per month to xAI for access to the Colossus 1 and Colossus 2 data centers, and it has a separate $200 billion, five-year commitment with Google Cloud for TPUs. The xAI deal alone could total more than $40 billion through May 2029.

At first glance, those numbers look incompatible with $559 million in quarterly operating profit. The way the math works out: most of the compute spend is amortized across longer time horizons, and a meaningful portion is treated as capital commitment rather than current-period cost. The deals secure capacity for future training runs, not just current inference workloads.

Anthropic itself has cautioned that full-year 2026 profitability is not assured. The company expects planned increases in compute and training spend through the second half of the year. Q2 may be the high-water mark of the year on margins, not the new baseline.

The funding round and the valuation

The Q2 numbers were shared with investors as part of Anthropic's ongoing funding round, which earlier reporting put at roughly $50 billion raised at a $900 billion valuation. Several sources have suggested the round could push Anthropic's valuation above OpenAI's, which would be a first in the modern AI era.

Whether or not the valuation comparison holds, the funding round is now happening with profitability as part of the pitch. That changes the type of investor likely to lead the round. Pure growth funds tend to lose interest in companies once profitability shows up; long-horizon institutional capital and sovereign wealth funds tend to like it. Expect a different mix of names on the final term sheet than what was rumored earlier.

Industry implications

For OpenAI, this is a competitive problem on multiple fronts. OpenAI's own revenue is large — reporting from earlier this year put the company at $25 billion in annual run-rate — but its profitability picture is more complicated, with heavier consumer subscription mix and a higher cost structure on inference. If Anthropic prints a profit while OpenAI prints a loss for the same quarter, that's the kind of split investors notice.

For enterprise buyers, Anthropic's profitability is a procurement asset. CIOs who have been asked to justify long-term commitments to AI vendors now have a cleaner answer when their boards ask about vendor risk. Profitable companies survive recessions; unprofitable ones get acquired or wound down.

For the broader AI industry, the result is a credibility marker for the enterprise-first business model. Anthropic has spent two years arguing that selling to businesses rather than consumers is the right strategy. The Q2 figures are the strongest evidence for that argument the company has produced.

What insiders are saying

The investor community's reaction has been to immediately revise its mental model of what late-stage AI companies look like. The assumption for years was that frontier labs were structurally unprofitable because compute costs scaled faster than revenue. Anthropic's numbers suggest that assumption was wrong, at least for one company at one moment in time.

Several VCs on X noted that the growth-vs-profit tradeoff in AI may now look more like enterprise SaaS than consumer internet. SaaS companies historically achieved profitability at lower revenue thresholds because their unit economics improve with scale. AI companies, the thinking goes, may follow a similar curve if they sell into enterprise rather than consumer.

What's next

Anthropic has not committed to releasing audited Q2 financials publicly. The figures shared with investors are projections, not closed quarters. Confirmation will come either in the next funding round announcement, where a more complete financial picture is likely to be disclosed, or in any eventual public filing if the company moves toward an IPO.

In the meantime, expect competitors to scrutinize the math. The compute commitments are public; the revenue figures are not yet audited. Skeptics will ask how the two reconcile.

The bottom line

For two years, the AI industry's biggest open question has been whether frontier labs can ever turn a profit. Anthropic just answered yes — or at least, yes for one quarter. Whether the answer holds across a full year is the next question. But the precedent is set, and the AI economics conversation will not be the same afterward.