OpenAI's $300 billion valuation works out to roughly 23x its ~$13B in annualized revenue — a price set in a SoftBank-led $40B round, the largest private financing in tech history. That's the short answer. The longer answer is more interesting.
A 23x revenue multiple on a company burning billions a year sounds insane until you separate what investors are paying for from what the company earns today. Nobody underwriting OpenAI at $300B is valuing the 2026 income statement. They are valuing a revenue curve that has gone from $3.7B to ~$13B in roughly eighteen months, and a product surface — ChatGPT — that became one of the fastest-adopted consumer technologies ever built. Whether that's genius or mania is the whole debate.
OpenAI Valuation $300 Billion: How the Number Was Set
OpenAI's $300 billion valuation was set in a primary round led by SoftBank in early 2025, with a commitment of up to $40 billion — the largest private tech financing on record. On roughly $13B in annualized revenue, that implies a forward revenue multiple of about 23x. The price is not anchored to current profitability, which is negative; it is anchored to the expectation that revenue keeps compounding 2-3x a year while inference costs fall.
To put the climb in context: OpenAI was valued at about $29B in early 2023, ~$86B in late 2023's tender, $157B in October 2024, and then ~$300B in the 2025 round. That is a roughly 10x increase in valuation over two years, tracking — and slightly outrunning — its revenue. The market is repricing the company faster than it is repricing the revenue, which is the definition of multiple expansion.
The one-line version
OpenAI is priced like the default interface to AI for a billion-plus people — not like a software vendor selling seats. The multiple is the cost of that assumption.
OpenAI's Valuation History and Revenue Multiple Over Time
The clearest way to understand how OpenAI is priced is to lay the rounds next to the revenue. The multiple has stayed in a 20-30x band even as the absolute numbers exploded — the market keeps paying roughly the same premium per dollar of revenue, just on a far bigger base.
| Date | Valuation | Annualized Revenue | Approx. Multiple | Lead |
|---|---|---|---|---|
| Early 2023 | $29B | ~$1B run-rate | ~29x | Thrive Capital |
| Late 2023 | $86B | ~$2B | ~43x | Tender (employees) |
| Oct 2024 | $157B | ~$4B | ~39x | Thrive Capital |
| Early 2025 | $300B | ~$10-11B | ~28x | SoftBank ($40B) |
| Mid 2026 | ~$300B+ | ~$13B | ~23x | Secondary / tender |
| Implied 2027E | (target) | ~$25-30B | ~10-12x | If growth holds |
Revenue figures are annualized run-rate estimates from public reporting; multiples are approximate. The 2027 row is illustrative of what bulls underwrite, not a disclosed figure. Compare these marks against the broader AI cohort on the AI Valuations dashboard.
Where OpenAI's $13B in Revenue Actually Comes From
The valuation only makes sense if the revenue base is durable, so it's worth breaking down. Unlike most foundation-model labs, OpenAI built a consumer business that now dominates its P&L — ChatGPT is the engine, not the API.
ChatGPT Plus / Pro (consumer)
~20M+ paying subscribers at $20-200/mo; the single largest revenue line
ChatGPT Enterprise / Team
Seat-based business plans; reported 1M+ business customers and fast-growing
API platform
Developer usage billed per token; high-volume but lower margin and more competitive
Microsoft revenue share
Azure OpenAI Service distribution; economics shared under the partnership
Roughly 70% of revenue is consumer and enterprise subscription, which is the part investors love: it's recurring, it's sticky, and it carries pricing power that pure API resale does not. The API business is bigger in mindshare than in dollars, and it's where margin pressure is most intense as Anthropic, Google, and open models compete on price. A subscription-heavy mix is exactly what lets a 23x multiple stand up — it looks more like a consumer platform than a commodity inference vendor.
How OpenAI Is Priced vs Other AI Leaders
A $300B price means nothing in isolation. Set against the rest of the AI landscape, OpenAI's multiple is high but not the highest — the pre-revenue frontier labs are priced far more aggressively per dollar of revenue.
| Company | Valuation | Annualized Revenue | Revenue Multiple |
|---|---|---|---|
| OpenAI | ~$300B | ~$13B | ~23x |
| Anthropic | ~$60-180B | ~$4-9B | ~20-25x |
| xAI | ~$50B+ | ~$1-2B | ~30-50x |
| Databricks | ~$62B | ~$3B | ~20x |
| SpaceX (for scale) | ~$350B | ~$15B | ~23x |
| Public SaaS median | — | — | ~6-8x |
The takeaway: OpenAI is priced like the category leader, not the category lottery ticket. Its ~23x is roughly 3x the public SaaS median but well below the 30-50x on companies with a fraction of the revenue. In other words, the market is paying up for OpenAI's scale and brand, but it is paying even more — on a relative basis — for unproven frontier bets. That's a useful tell about where the speculation actually lives.
The Risks Buried in the $300B Price
A 23x multiple is a promise that the future is much bigger than the present. Three things could break that promise.
What Holds the Valuation Up
- ✓ Revenue ~3.5x in 18 months ($3.7B → ~$13B)
- ✓ ~70% recurring subscription mix with pricing power
- ✓ ChatGPT brand as the default consumer AI surface
- ✓ Falling inference costs improving unit economics
What Could Break It
- ✕ Multi-billion annual losses on compute and talent
- ✕ Model commoditization as rivals match capability
- ✕ Microsoft dependency and AGI clause complications
- ✕ Capex arms race requiring perpetual fundraising
The biggest single risk is the cash burn. OpenAI is reportedly spending well beyond its revenue on compute and the Stargate-scale infrastructure buildout, which means the $40B SoftBank commitment isn't a war chest — it's closer to runway. A company that must keep raising at ever-higher prices to fund its losses is only one disappointing fundraising cycle away from a down round. The valuation assumes the financing window never closes.
What the $300B Price Tells Founders and Investors
I've made 65+ investments, and the OpenAI mark is the single most important comp in private tech right now because everything reprices off it. When the category leader is at 23x on real revenue, every Series A AI deck quietly assumes its own version of that curve. Most won't get it. The lesson isn't "AI is overvalued" — it's that the multiple is a function of revenue durability and growth rate, and almost nobody else has OpenAI's combination of both.
For founders raising in this market: the comp that matters isn't OpenAI's valuation, it's OpenAI's growth rate. If you can't show a credible path to compounding revenue 2-3x a year with a defensible mix, a 20x multiple is fantasy regardless of how hot the sector is. Track how the rest of the cohort is being marked on the AI Valuations and Unicorn Tracker dashboards before you anchor on any single headline number.
$300B isn't a price on what OpenAI earns. It's a price on what it becomes.
At 23x revenue, the market is underwriting the next $100B — not the last $13B.
Track AI company valuations and revenue multiples on the AI Valuations Dashboard at Value Add VC. Originally published in the Trace Cohen newsletter.