Bloomberg Reports OpenAI's Annualized Revenue Run Rate Passed $40 Billion; IPO Date Still Unset
Bloomberg, citing an internal memo from president Greg Brockman, says OpenAI's revenue pace has roughly doubled since the end of 2025, while reporting since late June says the company's planned listing may slip past 2026.
A Number, A Memo, and a Question Nobody Can Answer Yet
OpenAI's revenue is growing fast enough that even the company's critics don't dispute it. Bloomberg reported on August 13, 2026 that the company's annualized revenue run rate has passed $40 billion, citing people familiar with the matter and an internal memo from co-founder and president Greg Brockman[1]. Brockman told staff the monthly run rate grew more than 20% in July alone[2]. That's roughly double the pace CFO Sarah Friar described at the end of 2025, when she put the run rate above $20 billion[2].
A "run rate" is not the same thing as a year of actual sales. It takes whatever the company brought in during its most recent month and multiplies that by twelve, as if every future month repeated it exactly. So $40 billion isn't money OpenAI has collected. It's a projection built from one recent month, and it can move just as fast in the other direction if growth slows. Bloomberg credited the July jump to AI coding software, subscription sales, and a young advertising business — three separate revenue streams, not one product carrying the number[1].
Here's the tension: this same number is being read as proof the IPO is close, and as proof it might not be. Both readings use the same $40 billion. Neither is wrong about the figure. They disagree about what it means.
Two Reports, Two Timelines
Bloomberg's August headline framed the run rate as arriving "ahead of a targeted September IPO." But that timeline runs into a separate, earlier report. In late June 2026, Reuters said OpenAI was leaning toward waiting until 2027 to go public, and that Friar had told associates 2027 was the real target[5]. CNBC reported the same thing and noted the market reaction it triggered[5].
That reaction was not small. SoftBank, an OpenAI investor, saw its shares fall more than 12% in a single day on the delay report — its steepest drop in over three months[5]. That's a real market reading the delay as real news, not noise.
Nothing since has resolved which timeline is current. OpenAI has filed a confidential draft registration statement with the Securities and Exchange Commission, and Goldman Sachs and Morgan Stanley are reported as lead underwriters[6]. But a confidential draft is not a public filing — it doesn't require the company to release audited financial statements, and no price, date, or exchange has been announced[6][7]. Every number the public has seen, including the $40 billion, has come from a leak or a company statement, not a filed document[6].
There's also a plausible reason for OpenAI not to rush. In August 2026 the company completed roughly $7 billion in buybacks of employee shares at an $852 billion reference valuation[7]. A company that just handed employees billions in liquidity at a near-trillion-dollar price doesn't have the same urgent need for IPO cash that a cash-strapped startup would.
Prediction markets are pricing the uncertainty directly. As of mid-August, Polymarket put the odds of an OpenAI listing by the end of 2026 at about 19%[9] — not likely, but not zero either.
What OpenAI Wants Readers to Take From This
OpenAI's strongest case rests on the shape of the growth, not just its size. The company says the gains aren't limited to consumer ChatGPT subscriptions — they're spreading into enterprise coding tools and a new advertising business, three distinct revenue lines instead of one[1]. Leadership frames its spending less as a warning sign and more as a bet already paying off: the company is buying compute now for demand it says already exists, similar to a utility building a power plant before new customers plug in.
On timing, the company's position is that it isn't desperate. It just returned billions of dollars to employees through a buyback priced near $852 billion, which relieves some of the pressure that usually pushes private companies toward a rushed listing[7]. Sam Altman has reportedly told advisers he won't lower a trillion-dollar valuation target just to list sooner, arguing that pricing shares below what private investors already pay would undersell the company[5].
Investors who want to see OpenAI go public read the same growth number as historic. Adding roughly $20 billion in annualized revenue in under a year is rare at any scale, and enterprise contracts tend to renew, which bulls treat as more durable than one-off consumer subscriptions. They also point out that heavy losses during a buildout phase aren't unusual — Amazon spent years unprofitable while building warehouses. Underwriters Goldman Sachs and Morgan Stanley, along with holders like SoftBank, have a direct financial interest in a completed offering[5][6].
The Other Side of the Ledger
Skeptics don't dispute the $40 billion. They dispute what it's worth as evidence. Their core objection is that a run rate is the easiest number for a private company to make look good — it annualizes one strong month and nobody outside the company can audit the underlying figures[6].
They point to the spending side of the ledger for contrast. Independent reporting by Ed Zitron put OpenAI's 2025 spending at around $34 billion against far smaller revenue, with losses climbing sharply[12]. Other reports put the projected 2026 loss near $14 billion[11]. Skeptics argue that a growing top line means less when the company is burning cash on that scale to produce it.
They also flag a structural wrinkle: some of OpenAI's largest customers are also its biggest financial backers. Nvidia, Microsoft, and Oracle all sell OpenAI the chips and cloud capacity it needs, while also holding investments tied to its success — a pattern Bloomberg's own reporting has mapped in detail[10]. Money moving in a loop between the same handful of companies, skeptics say, can inflate how healthy the whole picture looks. They compare it to vendor financing in the late-1990s telecom boom, when equipment makers lent their own customers the money to buy their equipment, and the resulting revenue proved hollow once the lending stopped.
The same week the $40 billion figure surfaced, a separate storyline was already running: senior departures. Former COO Brad Lightcap and revenue chief Denise Dresser both left OpenAI in 2026, part of a wider pattern of executive turnover[4]. CNBC quoted an outside commentator calling the pattern a "huge red flag" ahead of a possible listing[4]. The overlap in timing — a growth memo landing during a departures story — is itself part of what skeptics find notable.
What the Coverage Reveals About Who's Reading
The split in how outlets covered this story tracks pretty closely with what each outlet's audience already cares about. Bloomberg's own headline, "OpenAI's Revenue Run Rate Tops $40 Billion Ahead of IPO," puts the growth number first and treats the listing as settled context, even though its own network's June reporting suggested the timeline was anything but settled[1][5]. Reuters restated the figure with more caution, putting "Bloomberg reports" directly in its headline as a signal the claim was secondhand[2].
Retail-investor outlets leaned into the bullish read. Benzinga paired the revenue figure with Polymarket's IPO-timing odds, treating a betting market's price as if it carried the same evidential weight as a reported financial figure[9]. That pairing can hand readers a false sense of precision about when a listing will actually happen.
CNBC ran the opposite emphasis the same week, leading with the executive departures and burying the revenue growth entirely[4]. 24/7 Wall St. went further, headlining a piece "So Why Are Its Top Executives Sprinting for the Exit Before the IPO?" — a characterization framed as a question, which lets the piece argue toward a conclusion it never states outright as fact. Coverage outside the U.S., including India's TechStory, mostly reproduced Bloomberg's framing without independent reporting or added skepticism[8].
What's Actually Left Open
Strip away the framing on both sides and what's left is two facts that aren't actually in conflict, whatever the headlines imply: OpenAI's revenue is growing fast, and OpenAI is losing a lot of money doing it[1][2][11]. Nobody serious disputes either one. The disagreement is about which of those two facts predicts what happens over the next few years — and that's a judgment call, not something the $40 billion figure by itself can settle.
The IPO timeline is even less settled than the revenue question. A confidential draft filing is on record, underwriters are hired, and OpenAI has not said when, or at what price, it plans to actually go public[6][7]. Until a real S-1 becomes public, everything the market is reacting to — the $40 billion included — is coming from leaks and internal memos, not from audited numbers anyone outside the company can check[6].
Summary
Bloomberg reported on August 13, 2026 that OpenAI's annualized revenue run rate has passed $40 billion[1]. That is roughly double the pace the company was on at the end of 2025, when Chief Financial Officer Sarah Friar said the run rate was above $20 billion[2]. The figure came from people familiar with the matter and from an internal note by co-founder and president Greg Brockman, who told staff the monthly run rate grew more than 20% in July alone[1][2]. Bloomberg credited growth in AI coding software, subscriptions, and a young advertising business[1].
A "run rate" is not annual revenue. It takes the most recent period's revenue and multiplies it out to a full year. So $40 billion is what OpenAI would collect over twelve months if the current month repeated twelve times. It is a forward projection dressed as a total, and it moves fast in both directions.
The headline claim that this comes "ahead of a targeted September IPO" is the most contested part. OpenAI filed a confidential draft registration statement with the U.S. Securities and Exchange Commission in mid-2026[6]. But Reuters reported in late June that the company was leaning toward waiting until 2027, with Friar telling associates the company was aiming for that year[5]. No pricing, date, or exchange has been made public. Prediction market Polymarket priced the odds of a listing by December 31, 2026 at about 19% as of mid-August[9].
The genuine dispute is not whether revenue grew. Everyone concedes it did. It is whether revenue growth of this kind tells you much about the company's value when the same company is reported to be losing billions and has committed to enormous compute spending. Reports put OpenAI's 2026 loss near $14 billion[11], and Bloomberg has documented a web of deals in which OpenAI's suppliers — Nvidia, Microsoft, Oracle — are also its investors[10]. Bulls say a business doubling at this scale is rare and worth paying up for. Skeptics say a run rate is the easiest number for a private company to make look good.
The Event
On August 13, 2026, Bloomberg reported that OpenAI's annualized revenue run rate had passed $40 billion, citing people familiar with the matter[1]. The report said co-founder and president Greg Brockman told staff in an internal note that the monthly run rate rose more than 20% in July[2]. Bloomberg attributed the acceleration to the company's AI coding software, subscription sales, and a new advertising business[1]. OpenAI has filed a confidential draft IPO registration with the SEC but has not announced a price, date, or listing venue[6][7].
Undisputed Facts
- Bloomberg reported the $40 billion-plus annualized run rate on August 13, 2026, sourced to people familiar with the matter, not to a public filing[1].
- CFO Sarah Friar has said OpenAI ended 2025 with an annualized run rate above $20 billion[2].
- Greg Brockman told employees in an internal note that the monthly revenue run rate grew more than 20% in July 2026[2].
- OpenAI filed a confidential draft registration statement with the SEC in mid-2026; Goldman Sachs and Morgan Stanley are reported as lead underwriters[6].
- OpenAI has not publicly announced an IPO date, price range, or exchange[7].
- Reuters reported in late June 2026 that OpenAI was considering delaying its listing to 2027, and that CFO Friar had told associates the company was aiming for 2027[5].
- OpenAI's most recent private valuation reference point is $852 billion, used in an August 2026 buyback of roughly $7 billion of employee shares[7].
- OpenAI is not profitable; multiple reports put its projected 2026 loss in the range of $14 billion[11].
- Several senior executives have left OpenAI in 2026, including former COO Brad Lightcap and revenue chief Denise Dresser[4].
The Pressure
Strip away the moralizing and blame. What structural realities persist regardless of which narrative wins?
- Compute bills come due on a schedule
- OpenAI has made very large multi-year commitments for chips and data centers[10]. Those payments are contractual and fixed. Revenue is not. That gap is the structural reason the company needs either continuous private funding or a public listing, regardless of how any quarter's growth reads[11].
- A private company controls its own numbers
- Until a public S-1 is filed with audited statements, OpenAI decides which figures reach the press and when[6]. A run rate is a legitimate internal metric, but releasing it through a leak, in a week when executive departures were the competing story, is a choice about attention.
- Suppliers are also shareholders
- Nvidia, Microsoft and Oracle both sell to OpenAI and hold economic exposure to it[10]. Each has a direct interest in OpenAI looking large and growing, because their own backlogs and stock prices are tied to it[10].
- Valuation discipline versus liquidity pressure
- Altman is reported to have refused a listing below a $1 trillion target[5]. Early investors and employees want liquidity sooner. The August buyback at an $852 billion reference price relieved some of that pressure without a listing[7].
Material realityOpenAI is collecting more money each month than it was a year ago, and the growth is real enough that suppliers, rivals and index investors are repricing around it[1][2]. It is also spending far more than it takes in, with reported 2026 losses near $14 billion[11]. Both things are true at once. No audited financial statement is public, so outsiders are reasoning from leaks and company statements[6]. The listing itself is not scheduled: a confidential draft is on file, underwriters are hired, and the date remains open, with reporting since late June pointing toward 2027[5][6]. Meanwhile the exposure is already spread through ordinary U.S. portfolios via Microsoft, Nvidia and Oracle, whose shares move on OpenAI news[10].
Narrative as a weaponOpenAI is the most active shaper here. A doubled run rate, delivered through an internal memo that reached Bloomberg, is the strongest available answer to a bad news cycle about executives leaving — it wants readers to believe the business is compounding faster than the churn matters. Underwriters and existing investors amplify that, because a higher perceived value sets the floor for any offering. Skeptics and short-side analysts push the opposite frame, wanting readers to see an unaudited projection propped up by circular deals among the same handful of firms. Financial trade press mostly transmits whichever number arrived most recently. The one thing readers should hold onto is that the growth figure and the loss figure are not competing claims — both sides largely accept both numbers, and disagree about which one predicts the next five years.
How Each Side Sees It
Each major actor’s view — how it frames things, its underlying incentive, and how it’s materially affected. Tap a side to read it.
Frames it asThe company's strongest case is that the growth rate is the story, not the loss. Revenue roughly doubled in about seven months[1][2]. The gains are broadening beyond consumer ChatGPT into coding tools, enterprise seats, and advertising — three separate businesses, not one lucky product[1]. Leadership argues the losses are chosen, not forced: OpenAI is buying compute today to serve demand that already exists, the way a utility builds a plant before customers plug in. On timing, its position is that it does not need the money right now. It just bought back about $7 billion in employee shares at an $852 billion reference price[7]. Altman has reportedly told advisers he will not cut the trillion-dollar target to list sooner[5] — the argument being that pricing an offering below what private buyers already pay would be selling the company short.
WhyAccess to public-market capital at a valuation that funds compute commitments without diluting control, and liquidity for employees and early investors[5][7].
Impact on themA strong revenue number raises the price OpenAI can defend in any listing and improves its terms with lenders and chip suppliers. A weak one does the reverse, while the compute bills stay fixed[10][11].
Frames it asThey argue that few companies in history have added roughly $20 billion of annualized revenue in under a year[1][2]. Enterprise adoption, they say, is the durable part: business software revenue tends to renew, so a dollar of enterprise run rate is worth more than a dollar of consumer subscription. They also point out that being unprofitable at this stage is normal for infrastructure-heavy companies — Amazon lost money for years while building warehouses. The relevant question, in their view, is gross margin per query and whether it improves as chips get cheaper, not this year's bottom line.
WhyUnderwriters Goldman Sachs and Morgan Stanley earn fees on a completed offering; existing holders such as SoftBank need a listing to mark and eventually exit their stakes[5][6].
Impact on themSoftBank shares fell more than 12% intraday on the late-June report of a delay — its largest such drop in more than three months — showing how directly listing news moves related stocks[5].
Frames it asTheir case is that "run rate" is the softest metric a private company can publish. It annualizes one good month and cannot be audited by outsiders. They point to the gap between revenue and spending: reporting by Ed Zitron put OpenAI's 2025 spending at about $34 billion against far smaller revenue, with losses up sharply[12]. They also object to circular financing — Nvidia invests in OpenAI, OpenAI buys Nvidia chips, and the same dollars show up as revenue in more than one place[10]. Their analogy is vendor financing in the late-1990s telecom boom, where equipment makers lent customers the money to buy their equipment and the revenue proved hollow. On top of that, they read the 2026 executive departures — Lightcap, Dresser, Fidji Simo, Kate Rouch — as a warning sign; CNBC quoted the pattern as a "huge red flag" before a listing[4].
WhyAnalytical credibility, and for short sellers, profit if AI-linked valuations fall[10].
Impact on themSkeptic pressure raises the disclosure bar OpenAI must clear in a public S-1, where audited numbers replace leaked ones[6].
Frames it asRetail buyers largely cannot own OpenAI today. Their interest is in whether a listing arrives, at what price, and how much is knowable beforehand. Their strongest argument is procedural: a confidential draft filing means the public sees no audited statements until close to the offering[6]. Until then, every figure — including $40 billion — reaches them through selective leaks that the company benefits from. They also carry indirect exposure already, through index funds holding Microsoft, Nvidia, and Oracle, whose results are tied to OpenAI demand.
WhyGetting a fair, informed shot at a large offering rather than buying the top from insiders who priced earlier[7].
Impact on themMicrosoft's stock fell 12% on January 29, 2026, erasing about $440 billion in market value, after it disclosed that 45% of its $625 billion cloud backlog was tied to OpenAI[10] — a reminder that ordinary index holders are already exposed.
Like this article?
The Bias Ledger average rating 4.2
The same story, as framed by outlets across the spectrum, ordered least to most biased. The bias score (1 = straight, 10 = heavily spun) is an AI assessment of that framing — click an outlet to see its track record. The tell is the word choice or omission that reveals the angle.
| Outlet | Vantage | Bias | How they frame it | The tell |
|---|---|---|---|---|
| Reuters | U.S./U.K. center, wire service | 2 | "OpenAI revenue run rate tops $40 billion, Bloomberg reports" — flat restatement, attribution in the headline. | Puts "Bloomberg reports" in the headline, which correctly signals the claim is second-hand. Adds little context on losses, so the number stands alone. |
| Bloomberg | U.S. center, financial-market focused | 3 | "OpenAI's Revenue Run Rate Tops $40 Billion Ahead of IPO" — growth number first, listing as the frame. | "Ahead of IPO" implies a scheduled event. Bloomberg's own industry reporting and Reuters' late-June story both indicate the timing is unsettled and may slip to 2027[5]. The story rests on anonymous sources and an internal memo, with no audited figure. |
| Benzinga | U.S. retail-investor trading media | 4 | "OpenAI Revenue Run Rate Tops $40 Billion Ahead Of IPO" — filed under prediction markets, paired with Polymarket odds. | Treats betting-market pricing as a comparable form of evidence to reported revenue. That framing gives readers a false sense of precision about the IPO date. |
| TechStory | Indian technology trade press | 4 | "OpenAI's annual revenue run rate crosses $40 Bn" — plain aggregation of the Bloomberg wire. | Calls it "annual revenue," softening the distinction between an annualized projection and money actually collected. Adds no local or independent sourcing. |
| CNBC | U.S. center, business news | 5 | "OpenAI talent exodus raises 'huge red flag' ahead of IPO" — same week, opposite emphasis. | Leads with a quoted alarm phrase in scare quotes, which lets the outlet carry the judgment without owning it. The revenue growth is not the lead here at all. |
| 24/7 Wall St. | U.S. retail-investor commentary | 7 | "OpenAI Just Hit a $40 Billion Run Rate. So Why Are Its Top Executives Sprinting for the Exit Before the IPO?" | "Sprinting for the exit" is a characterization, not a reported fact, and the rhetorical-question headline plants a conclusion the piece then argues toward. |
References
- OpenAI's Revenue Run Rate Tops $40 Billion Ahead of IPO — Bloomberg · U.S. center; subscription financial media owned by Michael Bloomberg
- OpenAI revenue run rate tops $40 billion, Bloomberg reports — Reuters · Center wire service owned by Thomson Reuters
- OpenAI's Revenue Run Rate Tops $40 Billion as IPO Nears — PYMNTS · U.S. payments-industry trade publication, advertiser-funded
- OpenAI talent exodus raises 'huge red flag' ahead of IPO — CNBC · U.S. center business news, owned by Comcast/NBCUniversal
- OpenAI is reportedly delaying its IPO. Here's when Kalshi traders think it will announce — CNBC · U.S. center business news, owned by Comcast/NBCUniversal
- OpenAI files for IPO as AI investment race intensifies — NBC News · U.S. center-left broadcast news, owned by Comcast
- OpenAI Stock & IPO 2026: Valuation, How to Invest, IPO Date — StartupHub.ai · Independent startup/AI trade site, ad-supported
- OpenAI's annual revenue run rate crosses $40 Bn — TechStory · Indian technology trade site, ad-supported aggregation
- OpenAI Revenue Run Rate Tops $40 Billion Ahead Of IPO — Benzinga · U.S. retail-trading media, revenue from brokerage advertising and data products
- AI Circular Deals: How Microsoft, OpenAI and Nvidia Keep Paying Each Other — Bloomberg · U.S. center; subscription financial media owned by Michael Bloomberg
- Facing $14B losses in 2026, OpenAI is now seeking $100B in funding. But can it ever turn a profit? — R&D World · U.S. research/industry trade publication, advertiser-funded
- Exclusive: OpenAI Losses Increased Nearly 8X in 2025, With Spending Hitting $34 Billion — Where's Your Ed At · Independent newsletter by Ed Zitron, an outspoken AI-industry critic; reader-funded