OpenAI vs. Anthropic: Who Will Win the AI Unicorn IPO Race?

On August 19, media reported that OpenAI's second-quarter revenue reached $6.7 billion, up 18% quarter-over-quarter, falling short of market expectations. Over the same period, Anthropic's preliminary revenue surpassed $11.5 billion, up approximately 14-fold year-over-year and over 140% quarter-over-quarter, surpassing OpenAI in quarterly revenue for the first time since its founding in 2021.

As of the end of July, Anthropic's annualized revenue run rate topped $65 billion, up more than sevenfold from the end of 2025. OpenAI still held a scale advantage in 2025, but the growth gap in 2026 is reversing this situation.

Even more critical is the divergence in their paths to profitability. OpenAI's second-quarter operating loss widened from $9.3 billion to $12.3 billion, incurring over $1.80 in operating loss for every $1 of revenue generated. Meanwhile, Anthropic turned its adjusted operating profit positive for the first time, with third-quarter GAAP operating profit expected to exceed $1 billion and its inference gross margin rising from 38% to 70%-85%.

Anthropic was able to overtake OpenAI within 18 months not by relying on superior model capabilities, but by tapping into an area OpenAI failed to cultivate deeply: developer workflows.

About 65% of OpenAI's revenue comes from consumer subscriptions. ChatGPT has over 200 million weekly active users; while the vast free user base forms a brand moat, converting them to paying users faces resistance, and the company still bears inference and computing costs for a large number of free users. The $20-a-month subscription model forces OpenAI to continuously invest resources into the general conversational experience rather than deep developer tools.

Between 75% and 85% of Anthropic's revenue comes from enterprise API calls, and without the burden of computing costs for a massive free user base, the company can concentrate its resources on its core product, Claude Code.

Claude Code is fundamentally different from OpenAI's Codex. Codex follows a Q&A model of "you describe requirements, I generate code," where humans remain the primary coders. In contrast, Claude Code embeds directly into the codebase, reading files, modifying code, and executing commands to complete long-chain software engineering tasks.

Developers only need to give natural language instructions, and Claude Code can autonomously complete the entire workflow from problem localization to test execution. Once integrated into the CI/CD pipeline, it transforms from an optional tool into an infrastructure component, with switching costs far higher than an individual user switching chat apps.

It is not that OpenAI lacks technical capability, but rather that its business model limits it from pushing in the same direction. Anthropic's technical barrier to entry lies not in model parameters, but in the workflow itself.

According to a report released by Ramp, a U.S. corporate spend management platform, as of July 2026, as many as 43.5% of U.S. companies were paying for Anthropic, while OpenAI's paid enterprise adoption rate was 39.7%, making Anthropic the new dominant player in the enterprise AI market.

Both companies have entered the IPO preparation stage, but their pacing differs noticeably. Anthropic confidentially submitted an S-1 filing to the SEC in June and is expected to list as early as September–October 2026. A $65 billion funding round completed in May brought its post-money valuation to $965 billion, surpassing OpenAI's $852 billion. Today, some investors' expected valuation for Anthropic has reached $2 trillion, which could surpass SpaceX to become the largest IPO in history.

OpenAI also submitted prospectus filings in June. However, CFO Sarah Friar stated at an all-hands meeting in August that the company "will become a public company in 2027."

Prediction market data indicates a 78% probability that OpenAI's listing will be delayed until 2027 or later. Fidelity Securities estimates its anchor valuation is closer to $700 billion to $800 billion.

In terms of listing timeline and recent growth rate, Anthropic holds a clear edge; in terms of brand recognition and user scale, OpenAI still retains an advantage. The capital market's final pricing will depend on three dimensions:

Sustainability of revenue growth. Anthropic's current growth rate far outpaces OpenAI's, but it relies heavily on the single-product breakout of Claude Code, and whether this growth can be sustained is key to its valuation.

The downward trajectory of infrastructure costs. Training models, procuring GPUs, and leasing data centers all require massive capital. The trends in gross margins and cash flows will determine the valuation ceiling for both companies.

The degree of valuation overstretch. Although Anthropic's official Series H valuation is $965 billion, recent over-the-counter trading has driven it up to nearly $1.5 trillion, with market expectations reaching as high as a $2 trillion IPO valuation—investors have clearly priced in extremely aggressive long-term growth assumptions in advance.

OpenAI's reliance on consumer markets, widening losses, and executive turmoil are suppressing its IPO valuation; meanwhile, Anthropic commands a higher premium thanks to its enterprise APIs, initial profitability, and a clearer listing timeline. Common challenges remain for both companies: rising computing costs, tightening enterprise budgets, and sustained pressure on profitability. Three variables—revenue growth rate, path to profitability, and governance structure—are reshaping market valuation logic for AI unicorns.

For investors, what truly warrants attention after listing is revenue growth, gross margin, cash burn, and infrastructure spending—these four metrics will determine who becomes the next-generation AI giant.

This content was translated using AI and reviewed for clarity. It is for informational purposes only.