OpenAI's $1 Trillion IPO: $25B Revenue, $57B Burn, and What It Means for Developers

OpenAI hit $25 billion in annualised revenue in February 2026. It is preparing for an IPO that could value the company at up to $1 trillion — which would make it the largest public offering in history, worth as much as Berkshire Hathaway. It is also burning through money at a rate that will reach $57 billion per year by 2027, and it does not expect to reach profitability until 2030. All three of these facts are simultaneously true, and understanding how requires understanding what OpenAI actually is right now.

The Revenue Number: $25 Billion and What It Represents

OpenAI crossed $25 billion in annualised revenue in February 2026, up from $21.4 billion at the end of 2025 — a 17% increase in approximately six weeks. This growth rate is not a typo. OpenAI is adding revenue faster than almost any company in history at this scale.

The revenue comes from three main sources. ChatGPT subscriptions — the $20/month Plus plan, the $200/month Pro plan, and team and enterprise tiers — are the largest consumer revenue stream. API access charges to developers building on GPT-4o, o3, and other models are the largest developer revenue stream. And enterprise contracts — large organisations paying for custom deployments, priority access, and compliance-grade service — are the fastest-growing segment.

Anthropic, by comparison, is approaching $19 billion in annualised revenue — a gap of approximately $6 billion that has narrowed significantly over 2025. Six months ago, OpenAI's revenue lead over Anthropic was substantially larger. The convergence is the story behind OpenAI's urgency to list: Anthropic's enterprise market share has grown from 50% of first-time enterprise AI spend in January 2026 to 73% as of March. OpenAI is winning on total revenue but losing on the enterprise segment that will define the next five years of AI commercial deployment.

The IPO: $1 Trillion Target, Largest in History

OpenAI is laying groundwork for an IPO that could value the company at up to $1 trillion. For reference: Apple is worth $3.4 trillion, Nvidia $2.8 trillion, Microsoft $2.9 trillion. A $1 trillion OpenAI would enter the S&P 500 as one of the ten most valuable companies in the world on day one of its listing.

The timeline is contested internally. Reuters and Bloomberg reporting suggests OpenAI is targeting a filing with the SEC as soon as the second half of 2026. CFO Sarah Friar has told associates the company is aiming for a 2027 listing. The gap between those positions reflects the tension between needing capital now and wanting to list at the highest possible valuation after more revenue growth.

OpenAI raised $110 billion in February 2026 at a pre-money valuation of approximately $730 billion — making it already, at a private valuation, one of the most valuable companies on earth. The IPO would convert that private valuation to a public market price, with Goldman Sachs and Morgan Stanley already in early discussions about underwriting roles.

At $1 trillion, the IPO would raise somewhere between $60 billion at the low end and substantially more at the upper range — capital that Sam Altman has described as necessary to fund the compute infrastructure requirements for AGI development, which he has projected at trillions of dollars through 2030.

The Burn Rate: $57 Billion a Year by 2027

OpenAI will not be profitable until 2030. That is the company's own internal projection. Between now and then, it needs to fund operations, compute, research, and infrastructure at an annual burn rate that will reach $57 billion by 2027.

To put that in context: $57 billion per year is more than the annual revenue of most Fortune 500 companies. OpenAI will spend $57 billion in a single year — on compute alone, a significant portion of that goes directly to Microsoft Azure and, increasingly, to AWS and Oracle — before it earns back its first dollar of profit.

The unit economics explanation: training and running frontier AI models is extraordinarily expensive. GPT-5.4 and o3 inference costs have come down significantly from GPT-4's launch in 2023, but the volume of queries has grown faster than the per-query cost has fallen. More users, more API calls, more enterprise workloads — total compute cost keeps rising even as per-token cost drops.

The IPO solves the burn problem in one way: public markets give OpenAI access to ongoing capital at scale without the friction of private funding rounds. A listed OpenAI can issue shares, take on public debt, and use its stock as acquisition currency — all of which are unavailable to a private company.

The Microsoft Restructuring: What Developers Need to Understand

The IPO preparation required OpenAI to restructure its relationship with Microsoft, and the terms of that restructuring matter for every developer building on OpenAI's APIs.

In October 2025, OpenAI completed its conversion from a capped-profit structure to a for-profit public benefit corporation (PBC). As part of that deal, Microsoft received a 27% stake in OpenAI PBC — down from an earlier reported position that gave Microsoft revenue-sharing rights until it recouped its $13 billion investment.

The restructuring reduced Microsoft's ongoing claim on OpenAI's revenue but gave it a permanent equity stake in the public company. For Microsoft, this is a bet that a $1 trillion OpenAI is worth more than the revenue-sharing arrangement — and given that OpenAI's public valuation will likely exceed the $13 billion investment by 70x, that calculation is correct.

For developers, the restructuring signals that OpenAI's API pricing is no longer directly coupled to Microsoft's revenue recovery needs. OpenAI can now make pricing decisions — including competitive reductions — based on its own commercial strategy rather than Microsoft's ROI requirements.

ChatGPT as Enterprise Productivity Tool: The IPO Pivot

The internal all-hands message that leaked before the IPO preparation went public was specific: ChatGPT must become a "productivity tool." CEO of Applications held a meeting with employees orienting the company "aggressively" toward high-value enterprise use cases.

This is the IPO framing decision. Consumer AI assistants trade at consumer multiples. Enterprise SaaS platforms trade at much higher revenue multiples. If OpenAI can credibly position ChatGPT as the enterprise productivity layer — the platform that replaces or integrates with Microsoft 365, Salesforce, Workday, and ServiceNow — it supports a $1 trillion valuation against $25 billion of revenue.

The math: $1 trillion / $25 billion = 40x revenue multiple. That is not insane for a high-growth enterprise SaaS company (Salesforce has traded at 8-10x; Snowflake at 20-30x at peak growth). But it requires the market to believe OpenAI is an enterprise software company, not a research lab that happens to sell API access.

The Frontier platform — now at the centre of the Microsoft lawsuit — is a core part of this repositioning. Frontier is enterprise software. Multi-agent workflows for HP, Intuit, Uber, State Farm. Recurring enterprise contracts. High switching costs. That is the business model that justifies a 40x revenue multiple.

What the OpenAI IPO Means for Developers

If OpenAI goes public, several things change for the developer ecosystem.

Pricing pressure increases. Public company OpenAI faces quarterly earnings calls. Revenue growth expectations from public markets will push OpenAI to monetise its API more aggressively. The current relatively developer-friendly API pricing — especially for GPT-4o Mini — may face upward pressure as the company optimises for revenue per token.

Open-source alternatives become more strategically important. An OpenAI optimising for public market revenue has different incentives than the research-mission organisation it was. Developers who depend entirely on OpenAI's API are exposed to pricing changes they cannot control. The hedge: build on open-source models (LLaMA, Qwen, Mistral) for cost-sensitive workloads and use OpenAI for genuinely frontier capability requirements.

The API product roadmap becomes more enterprise-focused. Features that serve large enterprise customers — compliance tooling, audit logs, role-based access, SLA guarantees — will receive disproportionate investment post-IPO. Developer-focused features that do not have a clear enterprise revenue path may slow down.

Anthropic becomes the credible alternative at enterprise scale. With 73% of first-time enterprise AI spend already, Anthropic's Claude is the realistic competitive alternative to OpenAI for enterprise developers who want to avoid vendor concentration risk in a post-IPO, public-market-optimised OpenAI.

Key Takeaways

• OpenAI hit $25B annualised revenue in February 2026 — up 17% in six weeks, driven by ChatGPT subscriptions, API access, and enterprise contracts
• IPO targeting up to $1 trillion — the largest public offering in history; filing targeted for H2 2026 or 2027; Goldman Sachs and Morgan Stanley in early underwriting discussions
• Burn rate reaches $57B/year by 2027 — OpenAI will not be profitable until 2030; the IPO solves the capital problem by giving access to public markets for ongoing fundraising
• Microsoft gets 27% of OpenAI PBC following the October 2025 restructuring; OpenAI's API pricing is now decoupled from Microsoft's revenue recovery requirements
• Anthropic is closing the gap: $19B annualised revenue vs OpenAI's $25B; 73% of first-time enterprise AI spend going to Anthropic vs 50% in January — OpenAI is losing enterprise ground while winning on total revenue
• Post-IPO: expect API price pressure, enterprise-focused roadmap, and reduced developer-friendly pricing as public market revenue expectations reshape OpenAI's product priorities
• The $1 trillion valuation requires a 40x revenue multiple — only justifiable if OpenAI is repositioned as enterprise software, not a research lab; Frontier and the ChatGPT enterprise pivot are the instruments of that repositioning