OpenAI Hits $25B Revenue and 900M Users — The IPO Math Is Getting Uncomfortable

OpenAI crossed $25 billion in annualized revenue in February 2026 — growing from $6 billion in 2024 and $20 billion at the end of 2025. ChatGPT now has more than 900 million weekly active users, approaching 1 billion. The company generates $2 billion in revenue every single month.

Those numbers are extraordinary. They are also the setup for the most complicated IPO in technology history.

The Growth Trajectory Is Genuinely Unprecedented

The speed of OpenAI's revenue growth has no precedent in enterprise software history. Most companies that reach $25 billion in annual revenue took 20-30 years to get there. OpenAI effectively did it in three:

• 2023: $2 billion in revenue
• 2024: $6 billion in revenue
• 2025: $20 billion in revenue
• Early 2026: $25 billion annualized run rate

That is a 3x year-over-year growth rate sustained across three consecutive years. For context: Salesforce took 17 years to reach $20 billion. Workday has never hit $10 billion. SAP took decades. The only meaningful comparison is Google's early advertising revenue growth — and Google was growing from a zero-marginal-cost product, not one that costs hundreds of millions in compute per month to operate.

The user numbers are equally striking. ChatGPT reached 100 million users in two months after launch in 2022 — the fastest consumer product adoption in internet history. It has since grown to 700 million weekly active users by July 2025, 800 million by October, and is now past 900 million approaching 1 billion. The growth has not plateaued.

The Valuation Math That Should Make Investors Nervous

At $852 billion post-money valuation and $25 billion annualized revenue, OpenAI trades at approximately 34x revenue.

For comparison:

• Microsoft: ~12x revenue
• Nvidia: ~25x revenue (after AI-driven surge)
• Google: ~7x revenue
• Meta: ~10x revenue
• Salesforce: ~8x revenue

OpenAI's 34x multiple is priced for perfection. It implies either that revenue will continue growing at 3x annually for several more years, or that the company will achieve profit margins far beyond any comparable software business, or both.

The complication is that OpenAI is not profitable. The company burns significant cash on compute, paying Microsoft hundreds of millions per quarter for Azure infrastructure. Exact figures vary by source, but multiple reports place OpenAI's 2025 operating loss in the $3-5 billion range on $20 billion in revenue — roughly a 15-25% operating loss margin. Revenue growing faster than costs is the bullish case. Compute costs scaling with model capability improvements is the risk.

The $122 billion funding round closed in April 2026 gives OpenAI enough runway to reach profitability without needing external capital — if the growth trajectory holds. SoftBank's structure on the round reportedly requires an IPO path, which is why public market valuation math matters now.

What $2 Billion Per Month Means

$2 billion in monthly revenue comes from several sources:

ChatGPT subscriptions: ChatGPT Plus at $20/month, ChatGPT Pro at $200/month, and ChatGPT Teams at $30/user/month. At 900 million weekly active users, even a 3-5% paid conversion rate generates 27-45 million subscribers. At blended pricing of $25/month average, that is $675 million to $1.125 billion in monthly subscription revenue alone.

API revenue: Developers and enterprises calling GPT-5.4, GPT-5.3, and earlier models through the API. Enterprise API contracts, which are typically annual commitments at significant volume, are growing fast. Goldman Sachs, Morgan Stanley, and dozens of major enterprises have announced Claude and GPT integration — OpenAI is winning the larger contracts.

Enterprise ChatGPT: ChatGPT Enterprise at custom pricing for large organizations. Microsoft's Copilot distributes GPT models at scale across Office 365, Teams, and Azure customers — Microsoft pays OpenAI a revenue share on that usage.

Sora and other products: Sora video generation, voice modes, and other premium features that monetize separately or as part of the Pro tier.

The mix matters for IPO pricing. High-margin API and enterprise revenue is worth more than lower-margin consumer subscription revenue. OpenAI has not published a revenue breakdown, which makes precise valuation difficult.

Anthropic's $19 Billion — The Gap That Tells the Real Story

Anthropic is approaching $19 billion in annualized revenue per reports from early April 2026. The gap between $25 billion (OpenAI) and $19 billion (Anthropic) is narrower than most people expected 18 months ago.

What is driving Anthropic's revenue:

• Claude 3.5 and Claude Sonnet 4.6 leading enterprise coding benchmarks — SWE-bench performance has made Claude the default choice for AI coding tools
• Claude Max and enterprise deployments with Goldman Sachs, legal firms, and healthcare organizations
• Claude for Life Sciences and Healthcare products (the Coefficient Bio acquisition reinforces this)
• The Anthropic API being the default in many agentic frameworks and multi-model pipelines

The competitive dynamic: OpenAI wins consumer (ChatGPT brand recognition, 900M users), Anthropic wins enterprise code (Claude's coding benchmarks). Both are growing at extraordinary rates. The AI market at $25B+ annualized revenue across just two companies is already larger than many traditional enterprise software categories.

What the IPO Timeline Looks Like

SoftBank's investment structure and the PBC conversion both point to an IPO in late 2026 or 2027. The mechanics:

1. PBC conversion must complete first — this is the legal transformation from nonprofit-controlled entity to Public Benefit Corporation. The IRS and California AG are both involved. Sam Altman's equity grant is contingent on this closing.

1. Sam Altman gets his equity — reportedly ~7% of the PBC, worth ~$60 billion at current valuation. This must be resolved before IPO filings.

1. Profitability path needed — public market investors will require a credible path to positive operating margins. At current revenue growth and compute cost trajectories, OpenAI models profitability in 2027-2028.

1. Regulatory clarity — the FTC and DOJ have both scrutinized AI company structure. Microsoft's 26.79% stake and the OpenAI-Microsoft commercial relationship will face antitrust review.

The most likely scenario: IPO in late 2027, at a valuation that reflects either continued 3x growth (implying $75+ billion in annual revenue by then) or a compression to more reasonable multiples (~15-20x revenue on $50+ billion, implying a $750 billion to $1 trillion IPO valuation).

What This Means for Developers

The revenue numbers have direct implications for the developer ecosystem:

API pricing: At $25 billion in revenue on a path toward profitability, OpenAI has less need to subsidise developer adoption through artificially low API prices. Expect GPT-5.x pricing to gradually increase as the consumer subscription base matures and enterprise contracts command higher rates. Use the LLM API Pricing Tracker to monitor rate changes in real time.

Model releases: $2 billion/month in revenue funds continued model development. GPT-5.5 (Spud) has completed pretraining. The revenue base supports the compute required to train progressively larger models and maintain multiple production tiers simultaneously.

Platform lock-in risk: At 900 million weekly active users and $25 billion in revenue, OpenAI is embedding itself into enterprise workflows at a scale that makes switching expensive. Developers building products on ChatGPT integrations or the OpenAI API face the same platform dependency risk as any developer who built on a dominant platform before its pricing power matured.

For the leaked cap table showing who owns what at $852 billion, see the OpenAI cap table breakdown. For the $122 billion funding round structure, see the OpenAI $122B funding analysis.

Key Takeaways

• $25 billion annualized revenue as of February 2026 — up from $6B in 2024 and $20B end of 2025, a 3x YoY growth rate sustained for three years
• 900 million+ weekly active users approaching 1 billion — the fastest consumer product growth in internet history from 100M users in 2022
• $2 billion per month in revenue from subscriptions (ChatGPT Plus/Pro/Teams), API, enterprise contracts, and Microsoft Copilot revenue share
• 34x revenue multiple at $852B valuation — priced for continued 3x annual growth; higher than Nvidia, Microsoft, or any comparable public tech company
• Not yet profitable: operating losses estimated at $3-5 billion on $20B 2025 revenue — compute costs at scale remain the core challenge
• Anthropic approaching $19B: gap narrower than expected — enterprise coding and life sciences driving Anthropic growth; OpenAI wins consumer, Anthropic winning enterprise dev
• IPO path: PBC conversion → Altman equity grant → profitability demonstration → IPO, likely late 2027 at $750B-$1T+ valuation