OpenAI Q1 2026: $5.7B Revenue, $3.7B Cash Burn, and a Path to a $1 Trillion IPO

The numbers that OpenAI has laid before prospective public market investors are simultaneously astonishing and alarming. In the first three months of 2026, the company generated $5.7 billion in revenue. It also burned through $3.7 billion in cash. Both figures tripled year over year. The trajectory is real. So is the gap between them.

OpenAI’s S-1 filing — the registration document submitted to the Securities and Exchange Commission ahead of what the company hopes will be a September 2026 IPO — provides the most detailed public look yet at the financial architecture of the world’s most consequential AI company. What it shows is a business growing at a rate that few technology companies have matched at anything close to this scale, while simultaneously consuming capital in quantities that raise genuine questions about when, if ever, the economics resolve.

The Revenue Story Is Real

$5.7 billion in a single quarter, tripling year-over-year, puts OpenAI on a roughly $20-22 billion annualized revenue pace heading into the second half of 2026. For context, when OpenAI hit $1 billion in annual recurring revenue in early 2023, it was considered a watershed moment for the AI industry. The company crossed $3 billion ARR by mid-2024, $10 billion ARR by early 2025, and has now roughly doubled that again.

The revenue base is diversified across three main streams: ChatGPT subscriptions (consumer and enterprise), API access for developers and enterprise customers, and enterprise licensing deals. ChatGPT, launched in November 2022, has grown into one of the most widely adopted software products in history, with hundreds of millions of users across more than 160 countries. The Plus, Team, and Enterprise subscription tiers have converted a significant fraction of that user base into paying customers.

The API business, which powers everything from startups building on GPT-5 to large enterprises integrating OpenAI models into their own products, has proven particularly durable. Unlike consumer subscription businesses, which are subject to churn and competitive pressure, enterprise API relationships tend to be sticky: companies build workflows and products around a model provider and face significant switching costs.

The Burn Rate Is Also Real

The $3.7 billion cash burn in Q1 is driven almost entirely by infrastructure. Chips, power, and data center costs account for the overwhelming majority of OpenAI’s cost base, and those costs are scaling nearly in lockstep with revenue. R&D spending reached $8.6 billion in the quarter — a figure that includes the cost of training frontier models, compensating some of the highest-paid researchers in the world, and maintaining the compute infrastructure required to run ongoing experiments.

Long-term commitments with Microsoft, Oracle, and Nvidia — for data center access, cloud computing, and GPU supply — mean that much of OpenAI’s cost structure is locked in years ahead. These are not expenses the company can quickly reduce if growth slows.

The net loss is striking in absolute terms: with $3.7 billion in cash burned against $5.7 billion in revenue, OpenAI’s effective operating margin remains deeply negative. The company has told prospective investors it does not expect to be cash-flow positive until approximately 2030. That four-year runway to profitability requires sustained revenue acceleration and, crucially, depends on the assumption that compute costs will eventually plateau or decline relative to revenue as model efficiency improves.

The Cash Cushion Buys Time

The most important number in the S-1 may not be the revenue or the loss — it is the $73 billion in cash and marketable securities sitting on OpenAI’s balance sheet, up from roughly $40 billion at year-end 2025. That war chest, accumulated through a series of funding rounds including last year’s $40 billion raise at an $852 billion valuation, provides approximately five years of runway at the current burn rate without any additional capital.

In practice, OpenAI expects to raise additional capital — including through the IPO itself — long before that runway expires. The September 2026 timeline for the public offering, if it proceeds as planned, could value the company at $1 trillion or more based on current market expectations. That would make it one of the largest technology IPOs in history and would add substantially to the available cash cushion.

The Competitive Context That Justifies the Burn

To understand why investors are willing to consider this valuation, it helps to understand the competitive framing that OpenAI has constructed. The company is explicitly positioning its investments as analogous to Amazon’s decision in the early 2000s to sacrifice near-term profitability to build AWS — a bet that the company that owns the foundational infrastructure layer of a new computing era will eventually command economics that more than justify the upfront cost.

The argument has some force. If the deployment of AI systems does represent a once-in-a-generation shift in how economic value is created — comparable in scope to the internet or mobile computing — then the company that establishes the dominant models, developer relationships, and enterprise integrations may be well-positioned for a long period of exceptional pricing power.

The counter-argument is also real. Unlike AWS, where Amazon controlled physical infrastructure that was inherently difficult to replicate, AI models are being reproduced and commoditized at a pace that raises genuine questions about durable competitive advantage. Anthropic, now valued at $965 billion, is competing directly on frontier model capabilities. Meta’s open-source Llama models are giving enterprises an alternative that requires no licensing fees. Google’s Gemini models are improving. Microsoft’s MAI family is closing gaps in coding. The moat may be narrower than the valuation implies.

What the IPO Means for the Industry

An OpenAI IPO at $1 trillion or above would be a defining event for the AI industry, not just for the company. It would set a public market precedent for how AI companies are valued, providing reference points that flow through to every private company in the ecosystem — including Anthropic, which filed its own S-1 documents shortly after OpenAI. It would give retail investors direct access to the AI buildout for the first time, democratizing exposure that has been concentrated in venture capital and sovereign wealth funds.

It would also force a new kind of accountability. Public companies must disclose quarterly financials, face analyst scrutiny, and respond to shareholder pressure in ways that private companies do not. For OpenAI, which has operated with an unusual and much-debated governance structure — a nonprofit board overseeing a capped-profit entity — the transition to public markets represents a fundamental shift in how the company is held accountable.

The financial disclosures in the S-1 have already surfaced facts that were previously opaque: the scale of the operating losses, the concentration of revenue in specific customer segments, the magnitude of long-term compute commitments. More will follow as the IPO process proceeds.

For now, the arithmetic is clear. OpenAI is generating more revenue than any AI company in history, burning more cash than any AI company in history, and sitting on more cash than most technology companies have ever accumulated. Whether that combination ultimately produces one of the greatest companies in the world, or one of the most expensive cautionary tales in venture capital history, depends on whether the next four years unfold the way the S-1 implies.