AI Titans Reverse Quarterly Reports! WSJ: OpenAI Q2 Revenue Up Only 18%, Far Below Anthropic’s Over 140% Growth

Recent quarterly data disclosed by The Wall Street Journal has sent shockwaves through Silicon Valley and Wall Street. OpenAI, once credited with defining generative AI, reported $6.7 billion in revenue for Q2 2026, up only 18% from Q1’s $5.7 billion, while operating losses ballooned from $9.3 billion to $12.3 billion. In contrast, long-perceived underdog Anthropic surpassed $11.5 billion in revenue during the same period, exceeding OpenAI in single-quarter revenue for the first time and achieving modest operating profitability.

This is not merely a fluctuation in performance—it signals a fundamental shift in how the AI industry evaluates value: from 'whose model is most famous' to 'whose unit economic model is most sustainable'.

OpenAI’s annualized revenue run rate (ARR) in Q2 2026 was approximately $40 billion. While an 18% quarterly growth rate might be acceptable for traditional software giants, in a period of accelerating AI adoption, it is interpreted as 'significant deceleration'. Worse, the company burns over $1.80 for every $1 earned, moving further from profitability and casting a shadow over its much-anticipated IPO narrative.

Anthropic, meanwhile, presents a different trajectory. Its Q2 2026 revenue exceeded $11.5 billion (up from $4.73 billion in Q1), with ARR surpassing $65 billion by end-July—over 70% growth from around $9 billion at the end of 2025. During this period, it achieved modest operating profit, turned its adjusted operating margin positive, and raised its inference gross margin to 70–85%.

According to U.S. enterprise LLM market research, as of June 2026, Anthropic’s paid enterprise penetration rate in the U.S. reached 34.4%, surpassing OpenAI’s 32.3%. Its enterprise market share stands at 32% versus OpenAI’s 25%.

Behind this divergence lies a split in commercial strategy. OpenAI’s revenue remains heavily reliant on ChatGPT consumer subscriptions and brand traffic, with a high proportion of free and low-cost users, creating friction in converting to enterprise clients. Additionally, Revenue Chief Denise Dresser, COO Brad Lightcap, and Fidji Simo—widely seen as Altman’s successor—departed within Q2, exacerbating governance instability and investor concerns.

Anthropic’s explosive growth engine is highly specific: its developer programming tool, Claude Code, enables Claude to read codebases and execute long-chain software engineering tasks, transforming model calls from 'single conversations' into 'continuous, high-token-consumption workflows'.

Approximately 75–85% of Anthropic’s revenue comes from enterprise-grade APIs. Claude Code reduces overall task-level costs with high accuracy, bypassing traditional IT procurement processes by infiltrating organizations 'bottom-up' via R&D budgets. This is precisely why its token efficiency outperforms competitors, allowing it to achieve profitability early in the scaling phase.

The capital markets’ shifting dynamics are equally evident. Anthropic has already secretly filed for an IPO, partnering with Morgan Stanley, Goldman Sachs, and JPMorgan Chase. Rumors suggest a listing as early as next month or October, with market expectations pricing its valuation near $2 trillion—potentially surpassing SpaceX’s $1.77 trillion record to become the largest IPO in history. Proceeds will target securing GPU/ASIC inference compute capacity, converting financial flexibility into production ceiling.

Despite higher name recognition and an ARR exceeding $40 billion, OpenAI’s widening losses and management reshuffling complicate the timeline of its confidential IPO filing. Some investors have begun focusing on whether Q3 growth can rebound after July’s new model launch.

The reversal in both companies’ quarterly reports marks a transition in AI competition—from the 'arms race of general-purpose large model parameters' to an 'efficiency race in vertical scenarios'. OpenAI proves that owning the most famous chatbot does not equate to having the healthiest business model; Anthropic demonstrates a new paradigm centered on developer productivity, task completion, and effective cost as moats.

For investors, the comparison of both firms’ financials sends a clear signal: in the AI infrastructure investment narrative, scale no longer automatically equals a moat. Token efficiency, organizational execution, and quantifiable scenario embedding are replacing 'largest model parameters' as the most critical competitive variables in the second half of the race. How the global AI landscape will solidify over the next two years depends on whether OpenAI can halt its bleeding through enterprise transformation and whether Anthropic can sustain Claude Code’s efficiency flywheel post-IPO.