OpenAI Revenue 2026: $20B ARR, $4B Monthly Run Rate, and the Path to Profitability

OpenAI hit a $20B annualized revenue run rate in mid-2026 — about $4B per month — with 700M weekly ChatGPT users, 20M paid subscribers, and a cash burn pace that projects to $115B in cumulative losses through 2029.

That's the short answer. The longer answer is more interesting, because growing revenue 5.4x in 18 months isn't the hard part for OpenAI anymore. The hard part is closing the gap between $20B in revenue and the $80B–$100B in annual compute commitments that come due over the next five years.

OpenAI Revenue 2026: $20B ARR Breakdown by Product

OpenAI's $20B annualized revenue run rate in mid-2026 breaks down into roughly $14B from ChatGPT subscriptions, $5B from API consumption, and $1B from Sora video plus licensing. ChatGPT Plus at $20/month is still the largest single product line, but Enterprise and Team seats grew the fastest — from $1B annualized at start of 2025 to over $6B by mid-2026.

Sources: OpenAI investor disclosures, The Information, FT, Bloomberg, reported leaks Q1–Q2 2026. Figures are annualized run rates, not GAAP recognized revenue.

How OpenAI Revenue Got From $1B to $20B in 30 Months

The clearest way to see what happened: OpenAI passed $1B annualized revenue in mid-2023, $2B by end of 2023, $3.7B for full-year 2024, $13B by end of 2025, and $20B by mid-2026. That's a compound monthly growth rate of roughly 9% sustained over 30 months — the kind of curve almost no SaaS company has ever produced at this revenue scale.

Three things drove the curve. First, GPT-5's mid-2025 launch lifted Plus conversion meaningfully — internal numbers suggest free-to-paid conversion went from roughly 2.4% to 3.1% after launch. Second, the $200/month Pro tier (launched late 2024) is now estimated at 500K+ subscribers — that's a $1.2B+ ARR line item alone. Third, Enterprise seat count crossed 4M by Q1 2026 across over 5,000 customers, with average contract value of about $800K and a top-end deal (Walmart, JPMorgan) above $50M annually.

OpenAI Burn Rate and the Path to Profitability

OpenAI is on pace for roughly $14B in 2026 net losses on $20B of revenue — a -70% operating margin. The internal financial plan, surfaced to the FT and The Information in late 2025, projects cumulative cash losses of about $115B through 2029 before the company turns cash-flow positive in 2029 at $125B annual revenue. For context, Amazon burned roughly $3B cumulative in its first decade. Uber burned about $25B before achieving GAAP profitability. OpenAI is in a different category entirely.

Three numbers explain the bulk of the burn. OpenAI's 2026 compute spend with Microsoft Azure is roughly $13B. The Stargate JV with SoftBank, Oracle, and MGX is in the early stages of a $500B 5-year buildout (about $100B/year run rate by 2027). Talent costs are running at roughly $4B annually on a headcount of approximately 4,500, or about $900K fully-loaded per employee — driven by competitive pay packages for AI researchers. Track this against the broader AI infrastructure spend on our AI Spending Dashboard.

OpenAI Revenue vs Anthropic, Google, and xAI in 2026

OpenAI is still the revenue leader among frontier AI labs in 2026 by a wide margin, but the gap to Anthropic shrunk dramatically. A year ago OpenAI had roughly 6x Anthropic's revenue. Today it's closer to 2.5x. xAI grew off a small base. Google Gemini's revenue isn't a clean comparable since it's bundled into Workspace and Cloud SKUs.

*Gemini revenue is estimated standalone; Google reports AI revenue inside Cloud and Workspace and doesn't break out Gemini-specific ARR.

OpenAI Revenue Per User and Unit Economics

With 700M weekly active users and $20B annualized revenue, OpenAI's blended revenue per weekly active user is about $28.50/year — but the distribution is wildly bimodal. The 680M free users generate roughly $0 in direct revenue (the ad product is still in test). The 20M paid subscribers generate an average of roughly $700/year — heavily skewed by Enterprise contracts.

• Plus tier ARPU: $240/year ($20/mo × 12) — about 11M subscribers — $2.6B run rate
• Pro tier ARPU: $2,400/year ($200/mo × 12) — about 500K subscribers — $1.2B run rate
• Team tier ARPU: ~$360/seat/year — about 2M seats — $720M run rate
• Enterprise ARPU: ~$800K/contract — about 5,000 customers, 4M seats — $4B run rate
• API ARPU: highly variable, top 100 customers do over $25M/year each — $5B run rate

Gross margin on ChatGPT Plus is reportedly around 50% — better than people assume, because the inference cost per query has dropped roughly 95% since GPT-4's launch in early 2023. API gross margin is closer to 70%. Enterprise contracts run lower because of dedicated capacity and SLAs but compensate with much higher ARPU and stickier multi-year terms.

What OpenAI Revenue Numbers Mean for Investors

At a $300B valuation against $20B ARR, OpenAI trades at a 15x revenue multiple — high but not absurd in 2026's AI market. The bull case is that the company hits its plan: $125B revenue by 2029, $175B by 2030, and a steady-state operating margin in the 30–40% range. At those numbers, $300B today looks cheap. The bear case is that compute costs scale roughly with revenue and never crack the 50% gross margin barrier, in which case OpenAI is a never-profitable utility.

For private market investors, the practical question is the secondary tender pricing. Thrive Capital, SoftBank, MGX, and Microsoft have all participated in tenders at the $300B mark. Employees can sell limited stakes at that price. Track this and the rest of the frontier AI valuation race on our AI Valuations dashboard.

For LPs in venture funds with OpenAI exposure (Sequoia, Thrive, Founders Fund, a16z, Khosla), the mark is a meaningful chunk of net TVPI on 2018–2021 vintage funds. Sequoia's OpenAI position alone was reported above $4B at the latest mark. Thrive Capital's is over $6B. These positions are real — but unrealized, and the path to DPI requires either an IPO (unlikely before 2027) or continued secondary tenders at or above current marks.