Everything About AI
NextFin News - OpenAI's annual recurring revenue is approaching $70 billion, with enterprise sales more than doubling since July, according to people familiar with the company's financials. The run rate has climbed more than 70% since the start of the third quarter, putting the ChatGPT maker within striking distance of rival Anthropic's roughly $65 billion annualized pace reported earlier this year - and setting the stage for what could become the most consequential artificial-intelligence initial public offering since the sector went public.
The acceleration matters because it reverses a narrative that had been hardening against OpenAI. For months, the market's central question was not whether artificial intelligence is growing, but whether OpenAI specifically was losing the enterprise race to Anthropic, whose Claude models had become the default for corporate and coding workloads. The latest figures suggest that question is being answered - and that the two labs are now converging on a combined annual revenue base that no software company in history has approached before going public.
The Numbers: A Three-Month Sprint That Rewrote the Growth Curve
OpenAI's annualized revenue run rate has grown more than 70% since the beginning of the third quarter, reaching almost $70 billion, according to the people with knowledge of the financials. The composition of that growth is as important as the headline figure:
- Business-to-business revenue more than doubled over the same period - the clearest sign that OpenAI is finally converting enterprise interest into contracted spend.
- On the consumer side, OpenAI added more revenue during the third quarter than it added during all of 2025 - a striking comparison given that 2025 was itself a record year for the company.
The trajectory is now visible in three markers. OpenAI ended 2025 with more than $20 billion in annualized revenue, according to Chief Financial Officer Sarah Friar. By mid-August, the run rate had passed $40 billion, roughly doubling from the prior year-end level. If the nearly $70 billion figure holds, OpenAI has added revenue in a single quarter at a pace that would take most mature software companies a decade.
There is one material caveat: details on OpenAI's expenses were not available, which is the missing half of the profitability picture. Leaked figures circulated earlier this year put OpenAI's 2025 net revenue at $13.1 billion against a net loss of $38.5 billion - a gap funded by the $110 billion capital raise the company closed in February at a $730 billion pre-money valuation, with SoftBank and Nvidia each committing $30 billion and Amazon adding a $50 billion stake for roughly a 5% share.
Why the Enterprise Turnaround Is the Real Story
The consumer business built OpenAI's brand; the enterprise business will determine whether it can justify a public-market valuation. Consumer subscriptions are high-churn, sentiment-driven revenue. Enterprise contracts are multi-year, embedded in customer workflows, and far harder to rip out once integrated into a company's software stack.
That is why the more-than-doubling of B2B revenue is the single most important data point here. It indicates that OpenAI's enterprise push - API integrations across finance, healthcare, and retail, plus ChatGPT Enterprise deployments - has moved from pilot programs to production contracts. Once an AI model sits inside a bank's underwriting pipeline or a hospital's documentation system, switching costs rise sharply. That is the difference between renting attention and owning a revenue stream.
The consumer acceleration is the second pillar. Adding more consumer revenue in one quarter than in the entire prior year suggests that new products - including the company's video-generation model and expanded subscription tiers - are monetizing a user base that has grown to more than 900 million weekly active users, rather than merely expanding the free tier.
The Anthropic Benchmark: Two Labs on Colliding Trajectories
The $70 billion figure cannot be read in isolation. Anthropic, the Claude maker, has been the market's growth champion this year. Its revenue grew twelvefold to nearly $4.6 billion in 2025, according to its IPO prospectus, and its run rate hit an annualized $65 billion in July before continuing to surge. The same prospectus disclosed $518 billion in future cloud, computing, and infrastructure obligations - a number that says as much about the capital intensity of this race as any revenue figure.
The two companies are now effectively racing toward the same finish line from different directions. Anthropic got to enterprise revenue first and owns the coding-agent category through Claude Code, which alone generated roughly $8 billion in annualized revenue by May. OpenAI arrived later to the enterprise but is growing faster from a larger installed base. If OpenAI's nearly $70 billion run rate is accurate and Anthropic's surge past $65 billion has continued, the two labs may now be separated by single-digit billions - a gap that could flip multiple times before either company lists.
The valuations tell a parallel story. Anthropic closed a $65 billion Series H round in May 2026 at a $965 billion post-money valuation, briefly surpassing OpenAI's roughly $852 billion mark. That ranking is itself a moving target: OpenAI's February round valued it at $730 billion pre-money, and private-market pricing has been volatile as investors reassess what a trillion-dollar AI company is worth when profitability remains years away.
The IPO Backdrop: A Public Market That Is Having Second Thoughts
The timing of this revenue report is not accidental. Both labs are preparing for IPOs that would give investors their first audited look at both the revenue opportunity and the extraordinary spending driving it. But the public-market window has narrowed.
On the same day this revenue figure surfaced, smart-ring maker Oura postponed its IPO, citing market uncertainty and denting hopes for a fall listings window. Investors are grappling with concerns that the AI trade has run its course, alongside higher bond yields and geopolitical turbulence. An AI IPO of OpenAI's scale would be the ultimate stress test of that sentiment.
OpenAI's own leadership has signaled caution. Chief Executive Sam Altman said in an interview that going public this year would be an "ill-advised moment," citing safety-related concerns, and indicated a listing would not happen until 2027. Rival Anthropic, by contrast, is discussing a public debut as soon as the fourth quarter of 2026, with bankers expecting a raise exceeding $60 billion and a valuation that some believe could reach $2.3 trillion.
"Right now would be an ill-advised moment to go public," Altman said, when asked directly about an offering that had been shaping up as one of the largest in market history.
That divergence in timing creates an unusual dynamic: Anthropic may set the public-market comps for the entire AI-lab sector before OpenAI ever files. If Anthropic's IPO prices well, it opens the door for OpenAI. If it stumbles, OpenAI's 2027 timeline starts to look less like caution and more like foresight.
The Hard Question: Is This Growth Structural or a Capital-Fueled Sprint?
Here is the judgment the market has not fully priced. OpenAI's revenue surge is real, but it is being produced by a business model in which the customer and the supplier are often the same balance sheet. OpenAI spends tens of billions on compute capacity - much of it from Microsoft Azure, in which Microsoft holds a 27% stake worth roughly $135 billion - and then recognizes revenue from the AI services that capacity delivers. A large share of the AI economy's revenue growth is, in effect, one hand paying the other.
That does not make the growth fake. It makes it capital-dependent. The structural question is whether OpenAI's revenue survives a slowdown in capex. If hyperscalers and enterprises keep building AI infrastructure regardless of short-term returns, OpenAI's revenue base is durable. If capital markets force a reckoning on AI spending - if the $518 billion in Anthropic-style infrastructure obligations starts to look unsustainable - revenue growth could decelerate as quickly as it accelerated.
The cyclical view says this is a capex supercycle: liquidity is abundant, every tech giant is racing for AI share, and revenue will mean-revert when the spending binge ends. The structural view says AI is a genuine platform shift - the way software is written, customer service is delivered, and scientific work is done is being rewritten - and that today's revenue is the first installment of a decade-long transition.
The evidence leans structural, but with a cyclical overlay. Enterprise contracts create retention; consumer monetization of a 900-million-user base creates optionality; and the fact that both labs can simultaneously grow revenue and burn tens of billions suggests demand is outstripping supply, not the reverse. But the margin structure is unproven at scale, and the loss figures - a $38.5 billion net loss on $13.1 billion of net revenue in 2025 for OpenAI, and a $42 billion net loss on $4.6 billion of revenue for Anthropic - show how far the economics still are from public-market expectations.
The Counter-Thesis: What If the Market Is Right to Hesitate?
The strongest argument against the bull case is simple: these numbers are run rates, not GAAP revenue, and run rates are the most optimistic possible framing of a business. They assume today's momentum persists, churn stays low, and customers do not renegotiate. Anthropic's own prospectus - the document that will govern how investors underwrite this entire sector - warns that its technology could pose "existential risk" to humanity, and a recent review found that top AI labs are investigating tens of thousands of incidents in which their systems took steps researchers would describe as problematic. A single major AI safety incident, or a regulatory intervention, could reset valuations overnight.
There is also the competition problem. Nvidia, the picks-and-shovels winner of this cycle, commands a $3.97 trillion market capitalization and posted $44.1 billion in quarterly revenue, up 69% year over year, with data-center sales of $39.1 billion. Microsoft sits at $3.2 trillion. These companies own the infrastructure layer and can, if they choose, move up the stack and compete directly with the labs they currently host. OpenAI's dependence on Microsoft Azure is both a partnership and a latent competitive threat.
The falsifying signal is quantifiable: if OpenAI's enterprise revenue growth slows below 30% year over year for two consecutive quarters while compute costs per token remain flat or rise, the structural-growth thesis breaks - it would mean the company is hitting demand limits rather than supply limits, and the revenue surge was a capital-fueled sprint, not a regime shift.
What Comes Next: Three Horizons to Watch
Short term (the next quarter): Watch whether OpenAI can sustain the third-quarter acceleration into the fourth. The consumer comparison - more revenue in one quarter than in all of 2025 - sets a high bar that will be difficult to repeat. A sequential slowdown would not invalidate the story, but it would confirm the cyclical overlay.
Medium term (the next 12 to 18 months): The IPO sequence matters more than any single revenue print. Anthropic's listing will establish the public-market multiple for AI-lab revenue. If it prices at a premium, OpenAI's delayed 2027 timeline becomes an advantage - it can watch, learn, and file into strength. If Anthropic stumbles, expect OpenAI to delay further.
Long term (three to five years): The winner will be the lab that converts revenue into margin. Today's race is measured in top-line growth because neither company is profitable at scale. The company that first demonstrates that AI revenue can be produced at software-like margins - rather than infrastructure-like costs - will capture the valuation premium. That is a race Anthropic, with its coding-agent focus and leaner cost structure, may be better positioned to win.
Base case: OpenAI files in 2027 at a valuation between $1 trillion and $1.5 trillion, with revenue above $100 billion - a target Chief Executive Sam Altman has hinted is reachable by 2027 - and a path to profitability still being underwritten by investors. Upside case: AI capital spending holds, enterprise adoption deepens, and OpenAI lists closer to $2 trillion alongside an Anthropic success. Downside case: capital markets force an AI spending reckoning, growth decelerates sharply, and both labs are forced to list at steep discounts to their private valuations.
The Bottom Line
OpenAI's dash to $70 billion in annualized revenue is a genuine acceleration, and the enterprise turnaround is the most credible evidence yet that AI is moving from experimentation into production budgets. But revenue without a profitability path is a growth story, not an investment thesis - and the public market is about to decide, through Anthropic's IPO, how much it is willing to pay for growth that is still being funded by the largest capital raise in technology history. The company racing to $70 billion may find that the harder number to hit is the one that comes after the dollar sign: the margin.
Explore more exclusive insights at nextfin.ai.
