OpenAI Valuation 2026: The $1.4T Round and Why There Is No IPO Yet

OpenAI’s last confirmed valuation is still $852 billion post-money, set when it closed a $122 billion round on March 31, 2026, and reaffirmed (not stepped up) by a company-funded employee tender in August 2026. A new round of about $30 billion at roughly $1.4 trillion was still only in talks as of October 19–22, 2026—not closed.

The arithmetic of the 2026 round is the cleanest way to read the source fight. OpenAI’s own February 27 blog set a $730 billion pre-money on $110 billion of commitments, which is an $840 billion post-money if that amount closed. When the round closed at $122 billion, post-money became $852 billion ($730B + $122B). The New York Times’s March 31 story still described the company as valued at $730 billion—that figure is the February pre-money, not the closing mark.

Chronological table

Figures are USD. “Confirmed” means a company, investor filing, or named primary announcement. “Reported” means press citing people familiar with the deal, often after the company declined to comment. Tender offers do not raise primary capital for the company unless noted.

Date	Event	Amount	Lead / buyers	Pre-money	Post-money / mark	How firm
Jul 22, 2019	Strategic investment (OpenAI LP)	$1B commitment, callable over years, not a lump sum	Microsoft	Not disclosed	Not disclosed	Amount confirmed by Microsoft and OpenAI. Parties declined to give terms or a valuation.
~Mar 2021	Follow-on Microsoft investment	~$2B (quiet; not announced at the time)	Microsoft	Not disclosed	Not disclosed	Reported, later supported by Fortune’s reading of deal documents and by court filings in the Musk litigation. No public valuation.
Jan 23, 2023	Extended Microsoft partnership	~$10B over multiple years (companies said only “multiyear, multibillion-dollar”)	Microsoft	Not stated	~$29B (reported target around the talks)	Amount reported (Bloomberg, NYT, person familiar). $29B reported via Semafor and later Fortune documents. Much of the value was Azure compute, not all cash.
Apr 28, 2023	Employee / insider share sale (tender)	Just over $300M	Sequoia, Andreessen Horowitz, Thrive, K2 Global; Founders Fund also reported; some accounts add Tiger Global	Not separately stated	$27–29B	Reported (TechCrunch documents; Reuters pickup). Separate from the Microsoft deal. OpenAI declined comment.
Feb 16, 2024	Employee tender	Amount not cleanly disclosed in primary coverage	Thrive Capital (lead)	Not stated	$80B or more (NYT) vs $86B (Bloomberg)	Reported. Both outlets said OpenAI and Thrive declined comment. Use $80–86B, not a single number.
Oct 2, 2024	Primary round (structured in part as convertible notes / interests)	$6.6B	Thrive lead (~$1.2–1.3B, with an option for more). Microsoft ~$750M to just under $1B. Nvidia ~$100M. SoftBank ~$500M. Also Khosla, Altimeter, Fidelity, MGX	~$150B+ (Bloomberg: more than $150B pre-money)	$157B	Round size and $157B valuation widely reported and treated as closed; OpenAI described the raise. Investor-by-investor splits are reported and disagree at the margin (Thrive $1.2B vs $1.25B vs $1.3B; Microsoft $750M vs “just under $1B”).
Nov 2024 (employees had until Dec 24 to elect)	Employee tender to SoftBank	Up to ~$1.5B	SoftBank Vision Fund 2	n/a (priced off the last round)	Same mark as Oct 2024 (~$157B; CNBC: $210/unit)	Reported (CNBC first; Reuters, WSJ, FT). FT said ~$150B, slightly looser than the $157B headline. Company and SoftBank declined comment.
Mar 31, 2025 (announced; first close planned mid-April)	Primary round, tranche structure	Up to $40B ($10B near-term; up to $30B later, cut if the for-profit conversion slipped)	SoftBank lead (effective up to $30B after syndicating $10B). Also Microsoft, Coatue, Altimeter, Thrive	$260B (SoftBank’s own announcement)	$300B including new money	Confirmed by OpenAI and by SoftBank’s April 1, 2025 filing for the pre-money, cap, and contingency. Syndicate names are reported.
Oct 2, 2025	Employee secondary sale	~$6.6B sold (authorization was ~$10.3B; only about two-thirds sold)	Thrive, SoftBank, Dragoneer, MGX, T. Rowe Price	n/a	$500B	Reported (Reuters, Bloomberg, NYT, CNBC). OpenAI did not confirm. Buyers largely declined comment. This is the mark Microsoft’s later 27% stake was discussed against.
Dec 11, 2025	Strategic equity investment	$1B, plus warrants	Disney	Not a new priced round	No new mark; described against the existing $500B secondary	Amount confirmed by Disney and OpenAI. Not a valuation reset.
Feb 27, 2026	Round announced, still open	$110B committed (Amazon up to $50B, of which $15B upfront and $35B milestone- or IPO/AGI-contingent; Nvidia $30B; SoftBank $30B)	Amazon, Nvidia, SoftBank anchors	$730B (OpenAI blog)	$840B if $110B counted in (Reuters, Axios, Guardian)	Pre-money and $110B confirmed by OpenAI. Post-money of $840B is arithmetic plus contemporaneous reporting, not a separate company sentence. Nikkei and some headlines used $730B as “the valuation,” which is the pre-money.
Mar 31, 2026	Same round closed, upsized	$122B committed	Anchors unchanged. SoftBank co-led with a16z, D.E. Shaw Ventures, MGX, TPG, and T. Rowe Price accounts. Microsoft participated (size undisclosed). $3B+ from individuals via bank channels	$730B (implied: 852 − 122)	$852B	Confirmed by OpenAI’s blog and @OpenAI post. CNBC, WSJ, Guardian, Bloomberg match $852B. NYT said $730B on the close day — that conflicts with the company and with the other wires.
Aug 10, 2026	Employee tender / buyback	~$7B	OpenAI itself (no new outside buyers)	n/a	$852B, flat	Reported (Bloomberg first; TechCrunch, CNBC follow-ons). OpenAI declined comment. First major tender that did not set a higher price.
Sep 29–Oct 19, 2026	New primary round in talks	About $30B	SoftBank, MGX, and other existing backers in talks; meetings with long-only and sovereign funds	Bloomberg (Sep 29): ~$1.4T not including the new money. Reuters (Oct 19): over $1.4T	Not set	Reported only. Not closed as of Oct 22, 2026. OpenAI and SoftBank declined comment to Reuters.

Related but not equity rounds: a roughly $4 billion credit facility around the October 2024 raise, later described by OpenAI in March 2026 as expanded to about $4.7 billion. That is debt capacity, not a valuation event.

Where the sources actually disagree

The fights are concentrated in a few cells, not spread evenly across the history.

• February 2024 tender. The New York Times said “$80 billion or more.” Bloomberg said $86 billion. Later round write-ups (WSJ, October 2024) used $86 billion as the prior mark. Treat it as a range.
• October 2024 check sizes. Everyone agrees on $6.6 billion and $157 billion. Thrive’s check is $1.2B, $1.25B, or $1.3B depending on the outlet; Microsoft is ~$750 million (Bloomberg) or just under $1 billion (WSJ). SoftBank ~$500 million and Nvidia ~$100 million come from people familiar, not a cap table.
• October 2024 pre vs post. $157 billion is the headline valuation. Bloomberg also said the deal valued the company at more than $150 billion before the new money, which is consistent with $157 billion post-money on a $6.6 billion raise.
• November 2024 tender price. CNBC tied the unit price to the last round ($210). The FT summarized the prior round as a $150 billion valuation. That is a rounding/pre-money slip, not a second price.
• March 2025. SoftBank’s own release is the cleanest primary: $260 billion pre-money, up to $40 billion, second tranche contingent on the recapitalization. OpenAI, Reuters, Bloomberg, and the FT all used $300 billion including the new money. “$40 billion” is a cap, not cash that all arrived on March 31.
• February–March 2026. This is the important one. OpenAI: $730 billion pre-money on February 27, then $852 billion post-money on $122 billion on March 31. Reuters/Axios/Guardian translated the February announcement into an $840 billion valuation including $110 billion. The New York Times on close day still wrote $730 billion. $730 + $122 = $852, so the Times figure is the pre-money carried forward, not a rival post-money.
• Amazon’s $50 billion is a commitment, not fully funded cash. OpenAI and Reuters: $15 billion first, $35 billion later if conditions are met (IPO or AGI, in Bloomberg’s account).
• October 2026 talks. Bloomberg on September 29: at least $30 billion at about $1.4 trillion excluding the new money. Reuters on October 19: about $30 billion at over $1.4 trillion, accelerated ahead of Anthropic’s expected mid-November IPO. Neither number is a closed round.

What is the latest mark?

Primary and company-confirmed: $852 billion post-money on March 31, 2026, on $122 billion of committed capital. OpenAI’s blog and its own post are the controlling sources. That implies a $730 billion pre-money, unchanged from the February announcement.

Last secondary print that matches a closed transaction: the August 10, 2026 employee tender, about $7 billion, priced flat at $852 billion, funded by OpenAI rather than by new investors. Bloomberg reported it; the company declined to comment. Flat tenders confirm a mark; they do not create a new one.

Not a valuation: the October 2026 fundraising talks. Until a round closes, “over $1.4 trillion” is a target described by people familiar with the matter, not a mark. Some secondary-data vendors (for example AG Dillon’s October 12, 2026 note) have already booked a $1.4 trillion “last primary” dated September 2026. That treats unsigned talks as a closed round. It should not be used as a confirmed valuation.

Secondary-market quotes, as of mid-to-late October 2026, do not agree with each other:

• Stock Analysis, citing Clarity (formerly Hiive), still lists the funding valuation at $852 billion (March 31, 2026) and an implied valuation of about $987 billion, with a last quoted price of $838.50 as of October 21, 2026.
• DefiLlama’s pre-IPO page, updated October 21, 2026, showed an estimated $1.11 trillion as of October 20, while still recording the last round at $852 billion.
• PrivateMarketView listed a secondary-based mark around $1.16–1.2 trillion (about $955 a share) as of October 2026.

Those marketplace marks are estimates from thin private trading, not transactions OpenAI has acknowledged. The spread—from the last closed print of $852 billion, through roughly $1.0–1.2 trillion in secondary estimates, up to a reported $1.4 trillion fundraising target—is the actual state of the mark as of October 22, 2026.

How to read the sequence if you are underwriting the next round

OpenAI’s valuation has been set by two different mechanisms, and they have not always moved together. From 2023 through October 2025, employee tenders did the repricing: about $29 billion (April 2023), $80–86 billion (February 2024), $157 billion was then locked by a primary round, $500 billion by the October 2025 secondary. The March 2026 primary jumped the mark to $852 billion; the August 2026 tender refused to raise it, and OpenAI paid for the liquidity itself. That is a different signal from 2024–25, when outside buyers (Thrive, then SoftBank, then a consortium) were willing to set a higher price to get stock.

The other structural point: several of the largest “investments” are commitments with conditions, not cash on the balance sheet. Microsoft’s 2019 billion was a capital commitment called over time. The 2023 Microsoft deal was multi-year and partly compute. SoftBank’s 2025 second tranche depended on the recapitalization. Amazon’s 2026 $35 billion depends on milestones. A headline raise of $122 billion is not the same object as $122 billion of unrestricted cash.

---

Recent Findings Supplement (October 2026)

OpenAI’s most recent confirmed primary round closed March 31, 2026, at a post-money valuation of $852 billion with $122 billion in committed capital. This round, anchored by Amazon (up to $50 billion), NVIDIA, and SoftBank alongside Microsoft and others, remains the last fully executed primary financing as of October 22, 2026.[1][2]

• OpenAI’s own announcement confirmed the $122 billion committed amount and $852 billion post-money valuation; major participants included Amazon, NVIDIA, SoftBank (co-lead), a16z, and others, plus over $3 billion via bank channels for individuals.[1]
• This built on prior rounds but marked a sharp step-up; committed capital typically arrives in tranches rather than immediate cash.[3]

This sets the baseline valuation for all subsequent liquidity and reported talks; any new primary round must clear or exceed this mark to attract capital amid rising infrastructure costs and IPO scrutiny.

SoftBank completed its final $10 billion tranche on October 1, 2026, bringing its total OpenAI investment to $64.6 billion for roughly 13% ownership. This followed the March round and was confirmed directly by SoftBank.[4]

• The tranche was part of SoftBank’s larger commitment in the $122 billion round; the firm raised bonds earlier to fund it.[4]

This demonstrates continued deep-pocketed backing from existing strategic investors even as new rounds are discussed, providing a signal of confidence in execution.

In August 2026, OpenAI completed a self-funded ~$7 billion employee tender offer at the unchanged $852 billion valuation, allowing current and former employees to sell shares directly to the company rather than outside investors. Bloomberg first reported the details, with confirmation across TechCrunch, CNBC, and others.[5][5]

• This differed from prior tenders (e.g., the October 2025 $6.6 billion round at ~$500 billion, which involved outside buyers); the company used its own cash to keep the cap table clean ahead of a potential IPO.[6]
• Earlier reporting referenced caps allowing some employees to sell up to $30 million each in prior liquidity events.[7]

This provides employee liquidity without diluting ownership or introducing new shareholders, a key pre-IPO tactic that preserves optionality while valuations remain high.

As of mid-to-late October 2026, OpenAI is in accelerated talks for a new ~$30 billion primary round targeting a valuation above $1.4 trillion, driven by competition with Anthropic’s expected mid-November IPO and revenue momentum. Reuters (Oct 19) and Bloomberg (late September updates) reported the details from sources; the round is not yet confirmed or closed.[2][8]

• SoftBank, MGX, and other existing backers are in talks; OpenAI met long-only funds and sovereign wealth funds recently.[9]
• Pitch includes September annualized revenue of ~$50 billion (down from some prior signals) with expectations of $70 billion by year-end, aided by GPT-6 Astra and enterprise growth.[10][11]
• Earlier reports mentioned possible $1.2 trillion targets; recent coverage converged on $1.4 trillion+ (pre- or post-money details vary slightly across sources, with one noting “not including the money raised”).[8][12]

Discrepancies center on exact valuation framing and timing; all figures remain reported rather than confirmed, with the round positioned as a bridge to a potential 2027 IPO amid market volatility.[3]

Secondary-market pricing as of October 21, 2026, showed shares at $838.50 (via Clarity platform), implying a valuation around $987 billion—above the March $852 billion mark but below the reported $1.4 trillion target.[13]

For competitors or new entrants, the combination of massive confirmed capital, self-funded liquidity tools, and aggressive reported valuation targets raises the bar for matching infrastructure spend and talent retention; overlapping investor pools with Anthropic intensify timing pressure on any IPO or raise. Recent revenue transparency and model launches (e.g., GPT-6 Astra) are being used to justify step-ups, while self-funded tenders signal balance-sheet strength but also highlight the need for sustained cash generation ahead of public markets.

The last closed OpenAI round is the March 31, 2026 financing: $122 billion of committed capital at an $852 billion post-money valuation. A separate ~$30 billion raise at roughly $1.4 trillion is still only in talks as of October 22, 2026, timed to lock capital before Anthropic’s expected mid-November IPO. The gap between those two marks is the whole story: primary capital is being asked to pay today for a year-end revenue target the company has not yet hit, while the last employee tender and the visible secondary tape still sit much closer to $852 billion–$1 trillion.

The closed mark: $122 billion at $852 billion post-money

OpenAI itself announced on March 31, 2026 that it had closed $122 billion of committed capital at a post-money valuation of $852 billion. That figure includes the new money. The February announcement that started the round was smaller and cleaner: $110 billion against a $730 billion pre-money valuation. The extra $12 billion, closed at the same price, lifts the arithmetic exactly: $730 billion pre-money plus $122 billion equals the $852 billion post-money OpenAI and Bloomberg both reported. The New York Times’s later reference to a $730 billion valuation was the February pre-money figure, not a second mark. [1] [2] [3] [4]

The round is a circular infrastructure financing dressed as equity. Amazon, Nvidia, and SoftBank supplied the overwhelming majority of the dollars, and two of those three also sell OpenAI the compute the capital is meant to buy.

• Amazon committed up to $50 billion. The first $15 billion was for Series C preferred; the remaining $35 billion was contingent on “specified milestones,” with an IPO or direct listing as the alternate trigger. Earlier reporting had framed the contingency as AGI or a listing. Amazon completed the full $50 billion by July 31, 2026, wiring a final $21.3 billion after June 30 once milestones were treated as met. In exchange, AWS became the exclusive third-party cloud distributor for OpenAI’s Frontier enterprise-agent platform, and the cloud contract was expanded. [3] [5] [6] [7]
• Nvidia committed $30 billion, tied to multi-gigawatt capacity on its systems (OpenAI has described 3 GW of dedicated inference and 2 GW of training on Vera Rubin). The Information reported that Nvidia paid a final $10 billion tranche on October 1 alongside SoftBank; that payment rests on anonymous sourcing and was not confirmed in a company release the way SoftBank’s was. [2] [8]
• SoftBank committed $30 billion in the follow-on, split into three equal $10 billion tranches executed April 1, July 1, and October 1, 2026, via Vision Fund 2. SoftBank’s own October 1 release confirms the final tranche is done. Cumulative investment is $64.6 billion for about 13% ownership. The first two tranches were drawn on a $40 billion bridge facility; the third was funded from an $11.1 billion foreign-currency high-yield bond sale announced September 24. SoftBank then canceled the remaining undrawn bridge capacity and said all bridge borrowings were repaid. [9] [10]
• Microsoft “continued” as a participant rather than as an anchor of new cash. Its own fiscal 2026 Form 10-K, as reported by secondary summaries of that filing, puts its as-converted interest at about 25% as of June 30, 2026, down from about 27% right after the October 2025 recapitalization, because later funding diluted it. Microsoft has funded about $11.9 billion of a $13 billion historical commitment. The nonprofit OpenAI Foundation retained control, with roughly 26% at the recapitalization. [1] [11] [12]

The rest of the syndicate is a who’s-who of crossover and sovereign capital, not a new strategic buyer. SoftBank co-led alongside a16z, D. E. Shaw Ventures, Abu Dhabi’s MGX, TPG, and accounts advised by T. Rowe Price. Other named participants include Altimeter, Appaloosa, ARK Invest, BlackRock affiliates, Blackstone, Coatue, D1, Dragoneer, Fidelity, Insight, Sequoia, Temasek, Thrive, and UC Investments. More than $3 billion came from individuals through bank channels—the first time OpenAI opened the round that way—and ARK said it would hold the shares in ETFs. An undrawn revolving credit facility was expanded to about $4.7 billion. [1] [1]

For anyone underwriting the next check, the mechanism matters more than the headline. Most of the $122 billion was not free cash from financial sponsors. It was strategic capital from chip and cloud vendors, partly milestone-gated, partly paid in tranches over seven months, and economically linked to purchase commitments that send a large share of the dollars back to the same investors.

The open mark: ~$30 billion at about $1.4 trillion, not closed

As of October 22, 2026, no new primary round has closed. Bloomberg reported on September 29 that OpenAI was aiming to raise at least $30 billion at a valuation of around $1.4 trillion not including the money raised—a pre-money figure—as a bridge after the company pushed its IPO out of 2026. Reuters reported on October 19 that OpenAI was racing to secure commitments for a round of about $30 billion that could value it at over $1.4 trillion, specifically before Anthropic’s expected mid-November listing. OpenAI and SoftBank declined to comment; the terms are unconfirmed. [13] [14] [15]

The investor list in the current talks is narrower and more familiar than the March syndicate. Reuters’s sources named SoftBank, Abu Dhabi’s MGX, and other existing backers, with meetings the prior week aimed at long-only funds and sovereign wealth funds. The point of the compressed timetable is competitive, not just operational: OpenAI and Anthropic are courting the same pool, and OpenAI wants a private mark set before Anthropic’s IPO disclosures become the public comparable. Edward Best, co-head of capital markets at Willkie Farr & Gallagher, told Reuters the logic is straightforward—set your own valuation before a listed peer forces the comparison. [14] [16]

There is a real pre- versus post-money ambiguity in the press. Bloomberg is explicit that $1.4 trillion excludes the new money. Reuters says “over $1.4 trillion” without the same precision. PitchBook’s October 12 note models a reported $1.43 trillion post-money figure. Those are not the same price. A $1.4 trillion pre-money on a $30 billion raise implies about $1.43 trillion post-money, which is why the two numbers keep appearing side by side. Until a term sheet is signed, both should be treated as reported asks, not marks. [13] [17]

The stated use of proceeds is balance-sheet insurance. Reuters’s sources said OpenAI needs the capital for infrastructure commitments and to fortify the balance sheet if a weaker IPO market pushes the planned 2027 listing further out. Sam Altman has said the company will not list in 2026 and has tied any debut to being able to make confident safety claims. Confidential S-1 work began in June 2026; the listing itself slipped. [14] [18]

Revenue basis: 17x on the last close, ~28x on the new ask

The multiple only makes sense once the revenue definition is fixed. OpenAI told investors its annualized revenue was approaching $50 billion at the end of September 2026, Bloomberg and the Financial Times reported in early October. A widely circulated ~$68–70 billion figure from late September was an investor construct that grossed up partner-channel sales so OpenAI could be compared with Anthropic, which counts the full value of contracts sold through Amazon, Microsoft, and Google. OpenAI’s own investor number counts only its share. The company is separately telling investors it expects annualized revenue to reach or exceed $70 billion by year-end, with the acceleration attributed in part to GPT-6 Astra and to price cuts on lower-cost models that lifted usage. At the March close it said it was generating $2 billion a month. [19] [20] [21] [14] [1]

That accounting choice moves the multiple by about 40%.

• Last confirmed primary: $852 billion post-money on ~$50 billion of late-September annualized revenue is about 17x. StockAnalysis lists a 17.0 price-to-sales ratio on a $50 billion run rate. On the $2 billion monthly figure disclosed at the March close, the same $852 billion was closer to the mid-30s on then-current run rate, so revenue has done a lot of the catch-up since spring. [22] [23]
• Proposed round: $1.4 trillion pre-money is about 28x the $50 billion OpenAI actually reported, and about 20x if the $70 billion year-end target is hit. PitchBook’s framing is the same math from the other side: a $1.43 trillion post-money is 28.6x the $50 billion run rate and 20.4x the $70 billion target. The round, in other words, prices December’s revenue in October. [20] [17]

PitchBook’s October 12 note puts fair value near $1 trillion, below the reported round, and notes that the year-end target assumes growth slows to roughly half the third-quarter pace. It also flags that costs are running at nearly three times revenue, so the multiple is not being earned by margin. Internal plans reported by the Financial Times have OpenAI burning on the order of $278–280 billion of cash from 2026 through 2030, with revenue in those projections rising from about $36 billion in 2026 to about $350 billion in 2030, and compute-and-infrastructure spending measured in the hundreds of billions. Those are press reports of company plans, not audited forecasts, but they are the numbers investors are being asked to underwrite. A 2025 operating loss of about $20.9 billion on roughly $13.1 billion of revenue has also been reported from the company’s accounts. [17] [24] [25] [26]

Secondaries and the tender sit well below the $1.4 trillion ask

The cleanest non-primary print of 2026 is flat. In August, OpenAI completed a roughly $7 billion employee share sale or buyback at the same $852 billion valuation as the March round. Unlike the October 2025 tender—about $6.6 billion at $500 billion, bought by outside investors including Thrive, SoftBank, and T. Rowe Price—the August deal was funded with the company’s own cash. No new names joined the cap table, and the step-up pattern that had run from $157 billion to $300 billion to $500 billion to $852 billion stopped. [27] [28] [29]

Live secondary quotes are higher than that tender, but not at the new ask. StockAnalysis, sourcing Clarity (formerly Hiive), showed a last price of $838.50 as of October 21, 2026, implying about $987 billion—roughly 16% above the March primary and the August tender, and about 30% below a $1.4 trillion pre-money. PrivateMarketView publishes a higher October mark of about $1.16–1.2 trillion, or roughly $955 a share, which it says is set by secondary prices and fund marks and is up 36% from the March round. Those two aggregators do not agree; Clarity’s quoted price and implied valuation are the more transparent of the two, and PMV’s share-level detail is largely paywalled. Either way, the observable secondary tape has not validated $1.4 trillion. [22] [30] [31]

The implication for a new investor is a three-tier market. Employees were cashed out in August at $852 billion. Accredited secondary buyers in late October are transacting in a band that implies something under $1 trillion on the clearest public quote, or low-$1 trillions on a less transparent aggregator. The company is asking primary investors for a pre-money nearly 65% above the last closed round and roughly 40% above the Clarity-implied secondary, in exchange for capital that existing strategic holders may or may not match.

Is the valuation justified? The argument, and the holes

The bull case is a distribution flywheel, not a model-quality argument. OpenAI’s March letter said it was generating $2 billion a month, had more than 900 million weekly active users and more than 50 million subscribers, and that enterprise was already over 40% of revenue and on track to match consumer by year-end. Later reporting put weekly users near one billion by August and annualized revenue near $50 billion by late September, with enterprise the growth engine. At 17x a $50 billion run rate, the March mark looks less extreme than it did in the spring, when revenue had not yet caught the valuation. If the year-end $70 billion target is real, even $1.4 trillion is about 20x forward run rate—rich for a company that is deeply unprofitable, but inside the range public software has paid for category leaders. [1] [19] [25]

The skeptical case is that the buyer set, the cash plan, and the comparable are all deteriorating at once.

• The multiple is being asked on revenue that has not arrived. PitchBook’s conclusion is that the round prices a target, and that fair value is nearer $1 trillion. The October 8 correction from a ~$70 billion narrative to a ~$50 billion reported run rate was large enough to hit the tape: Nasdaq and chip stocks sold off as investors repriced companies whose spending depends on OpenAI. [17] [20] [32]
• The capital is circular, so the valuation is partly a transfer price between a customer and its suppliers. Nvidia’s $30 billion and Amazon’s $50 billion sit next to multi-year chip and cloud purchase commitments. A meaningful fraction of “investment” is prepaid demand for the investors’ own products. That supports the infrastructure build; it does not independently validate what a financial buyer would pay for the equity. [2] [8] [6]
• The cash plan does not close without more raises. Press reports of OpenAI’s own multi-year plan show negative free cash flow on the order of $280 billion through 2030, with the March raise exhausted around 2028 and something like $100 billion-plus still to be raised even after this bridge. A $30 billion round, if it closes, is a waypoint, not a solution. [24] [25]
• The largest financial sponsor is already showing strain. SoftBank funded the last $10 billion with the world’s largest high-yield corporate bond issue of that moment, at yields approaching 10%, and its credit-default swaps were already elevated. After the Reuters report that SoftBank and MGX were in talks for the new round, SoftBank CDS jumped 12 basis points overnight to 432, a four-year high, according to market commentary. A 13% holder financing follow-on equity with junk debt is not a signal that the $1.4 trillion price is easy to clear. [10] [33]
• The comparable OpenAI is trying to front-run may not cooperate. Anthropic’s hoped-for IPO valuation has already been walked back in the press from about $2 trillion toward something nearer $1.5 trillion after weaker-than-hoped results, even as some investors still want a higher number. If Anthropic lists and the public market applies a lower multiple to a peer with disclosed losses and huge compute commitments, OpenAI’s private $1.4 trillion becomes the number that has to be defended, not the number that sets the market. [34] [18]

What this means for anyone trying to enter or compete with this cap table: the March round is a closed, documented fact at $852 billion post-money, mostly strategic money, heavily tranched and partly conditional. The $1.4 trillion figure is an unconfirmed pre-money ask on a round that had not closed by October 22. Secondary and tender pricing have not ratified it. The valuation is justified only if annualized revenue actually reaches ~$70 billion by year-end and public or crossover buyers accept ~20x that run rate on a business whose own plan still requires well over $100 billion of additional capital before it turns cash-flow positive. If either assumption fails, the last price that both employees and the company were willing to transact at—$852 billion in August—is the more relevant mark.

---

Recent Findings Supplement (October 2026)

OpenAI is accelerating talks for a ~$30 billion primary financing round targeting a valuation above $1.4 trillion (reports indicate this figure is pre-money or excludes new capital) as of mid-to-late October 2026, positioning it ahead of rival Anthropic’s expected mid-November IPO.[1][2]

This compressed timeline reflects competition for overlapping capital pools from sovereign wealth funds and institutions, alongside OpenAI’s recent revenue momentum. The round builds on earlier September/early October reports of similar size and valuation targets and serves as a potential bridge or pre-IPO financing while OpenAI plans a 2027 listing.[3][4]

• Size and structure: Approximately $30 billion in new commitments sought; no detailed tranche structure or conditions (e.g., milestones) publicly disclosed for this round. It follows the March 2026 close of a much larger $122 billion round at an $852 billion post-money valuation.[5]
• Investors: Existing backers SoftBank Group and Abu Dhabi-based MGX (a sovereign wealth fund) are in talks to participate; no confirmed new commitments from Nvidia, Microsoft, or additional sovereign funds in the latest reporting. SoftBank and MGX previously participated in the March round.[1][2]
• Valuation basis: Targeting >$1.4 trillion (pre-money per Bloomberg reporting relayed in multiple outlets), a substantial step-up from the March $852 billion post-money mark. OpenAI is pitching investors on an annualized revenue run-rate of ~$50 billion (post-GPT-6 Astra launch), with projections to reach $70 billion by year-end, implying a revenue multiple in the mid-to-high 20s at the target valuation.[4][3]

SoftBank completed the final $10 billion tranche of its $30 billion commitment from the March 2026 round on October 1, 2026 (funded partly via a record high-yield bond issuance), bringing its cumulative OpenAI investment to $64.6 billion for an approximate 13% stake.[6][7]

This closes out SoftBank’s three-tranche structure ($10 billion each in April, July, and October) tied to the prior round. Nvidia’s parallel $30 billion commitment from the same round appears fully deployed without similar tranching details in recent reports. Amazon’s up-to-$50 billion commitment included contingencies (portions tied to IPO or AGI milestones).[8]

Secondary-market and tender activity shows valuations in the $1.16–1.4 trillion range in October 2026, generally below or aligned with the proposed primary round target.[9]

• PrivateMarketView marks imply ~$1.16–1.2 trillion (~$955/share) based on secondary trading as of October 2026, up ~36% from the March primary round.[10]
• Other platforms (e.g., AG Dillon) reference ~$1.4 trillion implied valuations tied to recent primary discussions or marks.[11]
• An August 2026 employee tender/share sale of ~$7 billion occurred at the flat $852 billion valuation (company-funded, no new external capital or repricing).[12]

Analyst and market commentary frames the $1.4 trillion target as aggressive amid overlapping investor pools, rising scrutiny of AI infrastructure costs/valuations, and competitive dynamics with Anthropic (whose own IPO valuation expectations have faced downward pressure).[5]

OpenAI highlights post-GPT-6 growth and pricing flexibility to justify the step-up, while the timing aims to lock in capital before Anthropic disclosures enable direct comparisons. Broader context includes concerns over sustained losses (with profitability forecasts extending to 2030 in some plans) and massive compute commitments, though specific justification analyses for the exact multiple remain limited in available reporting.[2]

For competitors or entrants: The rapid valuation escalation (from $852 billion post-money in March to $1.4 trillion+ targets) underscores the premium on demonstrated revenue traction and strategic compute/investor ties, but also highlights execution risks around closing amid rival IPOs and valuation scrutiny. Overlapping investor bases and tranching precedents (e.g., SoftBank) suggest structured deals with contingencies may persist. Secondary liquidity at lower marks could pressure primary terms or provide alternative entry points.[13]

No major new regulatory, policy, or independent research findings on the financing itself appear in post-April 2026 sources; updates center on deal timing, investor participation, and revenue narratives. Details remain fluid as talks continue.

OpenAI’s late-September 2026 annualized revenue was about $50 billion on the company’s own books—not the ~$70 billion figure that circulated a week earlier. The gap is an accounting construction, not a $20 billion sales miss. The $70 billion number investors and outlets used in late September was built to match Anthropic’s gross treatment of cloud-partner sales. Separately, OpenAI has told investors it expects its own (net) annualized revenue to reach or exceed $70 billion by year-end 2026. That target is a forecast of exit pace, not recognized revenue for the calendar year, which internal plans still put near $36 billion.

The $50 billion vs. $70 billion reconciliation

On October 8, 2026, the Financial Times reported that OpenAI had told investors its annualized revenue was “approaching $50 billion” at the end of September. Bloomberg, CNBC, and Reuters confirmed the same figure. Nine days earlier, Axios had reported a run rate nearing $70 billion (some outlets cited about $68 billion). The Financial Times and a person familiar with the matter, as relayed by CNBC and Reuters, said that higher number came from OpenAI’s own investors trying to produce a like-for-like comparison with Anthropic—not from a company disclosure of $70 billion already achieved. [1] [2] [3] [4]

The mechanism is partner-channel accounting. Anthropic books the full customer payment on sales through cloud partners (AWS, Google Cloud) as revenue and records the partner’s cut as an expense. OpenAI, for certain partner sales—especially models sold through Microsoft Azure—recognizes only its own share. Reporting on the partnership, including CNBC citing a person familiar with the April 2026 renegotiation, puts that share at 20 percent, continuing through 2030 but now capped at a cumulative $38 billion. On a simplified $1 Azure sale, OpenAI may book roughly 20 cents while an Anthropic-style gross method books the full dollar. Secondary accounts of the investor math describe exactly this adjustment: take OpenAI’s net base and gross up the partner channel so the headline can be compared with Anthropic’s. That is how a ~$50 billion net figure became a ~$68–70 billion “comparable” figure. Neither company has published audited quarterly statements, and annualized revenue has no GAAP definition, so both approaches can be defended—and neither is cash in the bank. [5] [6] [7] [8]

A further wrinkle: after the April 2026 commercial update, OpenAI stopped receiving the revenue share Microsoft had been paying OpenAI on Azure resales of OpenAI models. Microsoft still collects its share of OpenAI’s own revenue. Mid-August reports of a run rate above $40 billion still included some of that inbound Microsoft revenue, which then dropped out of OpenAI’s recognized top line. The late-September investor “gross-up” and OpenAI’s own net figure are therefore not the same series. [9]

On October 9, Bloomberg reported a different $70 billion: OpenAI expects to reach or exceed $70 billion in annualized revenue by the end of 2026, on the basis it uses with investors, driven mainly by enterprise. That is a three-month forecast, not a restatement of September. Moving from roughly $50 billion to $70 billion is about a 40 percent increase in the run rate in one quarter. The same investor materials showed third-quarter run-rate growth of 77 percent overall and 107 percent in enterprise. [2] [10] [11]

What this means for anyone underwriting the company or its suppliers: quote the metric and the basis in the same sentence. A $1.4 trillion pre-money valuation discussion (Bloomberg, September–October 2026 fundraising talks for $30 billion or more) is about 28 times the September net run rate and about 20 times the investor gross-up. The multiple is not the story; the definition is. Suppliers are paid from the smaller number—the cash OpenAI actually keeps—not from a grossed-up comparison built for Anthropic.

Run rate is not recognized revenue

Annualized revenue multiplies a recent month (or short period) by 12. It measures pace. Recognized revenue is what the books record over a full period. Because growth is so steep, exit run rate has run far ahead of the year’s booked total.

• End of 2025: CFO Sarah Friar said annualized revenue had surpassed $20 billion, up from $6 billion in 2024 and $2 billion in 2023, as compute capacity rose from 0.6 GW to 1.9 GW. Audited full-year 2025 revenue was $13.07 billion. By December 2025 the company was generating nearly $2 billion a month—an exit pace near $24 billion against a $13 billion year. [12] [13] [14]
• First half of 2026: investors were told revenue was $5.7 billion in Q1 and $6.7 billion in Q2 (up 18 percent quarter on quarter), or about $12.4 billion booked in six months. Anthropic’s Q2 was reported at $11.5–11.6 billion, the first quarter it led on that metric. Those quarterly figures are recognized revenue, not run rates, and they are not adjusted onto a common partner-accounting basis. [15] [3]
• Full-year 2026 plan: the July presentation reviewed by the Financial Times projects about $36 billion of revenue for 2026, rising to $350 billion in 2030, with cumulative revenue of about $840 billion through 2030. Secondary summaries have also cited a Bloomberg figure near $35 billion for recognized 2026 revenue. Either way, a $50 billion September run rate and a $70 billion December target do not imply $50–70 billion of GAAP revenue in calendar 2026. [16] [17] [18]

Consumer versus enterprise is a mix shift, not a separate top line that explains the $20 billion gap. Axios reported that third-quarter consumer revenue alone exceeded all consumer revenue in 2025, while enterprise more than doubled since July. The growth rates OpenAI showed investors—77 percent total, 107 percent enterprise—are run-rate growth, not a split of the $50 billion. One mid-2026 account put consumer near 70 percent of revenue and enterprise above 40 percent; those shares can reflect different definitions and should not be added. The year-end $70 billion target is described as enterprise-led. [19] [20]

Cash burn and projected losses

The operational picture in 2025 is an operating loss, not the headline net loss. Audited figures obtained by Ed Zitron and reviewed by the Financial Times show 2025 revenue of $13.07 billion (from $3.7 billion in 2024), R&D of $19.18 billion, total costs of about $34 billion, and an operating loss of $20.92 billion (from $8.78 billion in 2024). Net loss attributable to the company was about $38.5 billion, swollen by a roughly $41.55 billion non-cash charge tied to the nonprofit-to-for-profit conversion. A person familiar with the matter told the FT that stripping that charge and other non-cash items left losses near $8 billion. Payments to Microsoft in 2025 totaled about $17.2 billion across R&D, cost of revenue, and other items. Year-end assets were just over $50 billion, roughly half in cash. [13] [21] [22]

Internal plans have been revised upward on both revenue and burn, then partially walked back on cumulative free cash flow:

• February 2026 (The Information): cash burn of $25 billion in 2026, $57 billion in 2027, $85 billion in 2028, $51 billion in 2029, and positive $39 billion in 2030. That vintage added about $111 billion of cumulative burn versus the prior forecast and assumed about $665 billion of training and inference cost through 2030. 2025 cash burn was about $8 billion. [23] [24]
• May 2026: negative free cash flow of about $305 billion over 2026–2030. [16]
• July 2026 presentation (FT, September 18): negative free cash flow of $278 billion from 2026 through 2030—better than May, still enormous—against $856 billion of computing and infrastructure spend and the revenue path from $36 billion to $350 billion. The $122 billion raised in March 2026, at an $852 billion valuation, is projected to be exhausted in 2028. [16] [17]

Quarterly losses were still widening into mid-2026. Investor disclosures reported by outlets covering the Wall Street Journal put the Q2 operating loss, including stock-based compensation, at $12.3 billion, up from $9.3 billion in Q1, while revenue rose only $1 billion. Losses grew faster than sales in that quarter. [25] [26]

The implication is a funding treadmill, not a near-term profit inflection. Even the improved July plan does not turn free cash flow positive before 2030, and the cash raised in March is modeled to run out two years earlier. A new round of $30 billion or more, if it closes, covers a fraction of the $278 billion five-year gap. Microsoft’s fiscal 2026 disclosure—$24.1 billion of revenue from commercial arrangements with OpenAI, including revenue share and Azure, and $6.0 billion of receivables at June 30—shows the cash is already flowing to the compute supplier, not sitting as OpenAI margin. [27] [28]

Compute commitments versus the revenue that has to pay them

Commitments are multi-year contractual or reported obligations. They are not annual P&L expense, they overlap, and several of the largest are press figures rather than line items OpenAI has confirmed in its own filings.

Disclosed or company-stated commitments that suppliers have put in their own documents or announcements:

• Microsoft: an incremental $250 billion of Azure services under the restructured agreement (Microsoft subsequent-event disclosure). Revenue-share payments to Microsoft continue through 2030, capped at $38 billion in total. [11] [7]
• Amazon: an existing $38 billion multi-year commitment expanded in early 2026 by $100 billion over eight years, including obligations tied to AWS chips—about $138 billion combined, per Amazon’s June 30, 2026 10-Q language as reported. [11] [29]
• CoreWeave: 2025 contracts of up to $11.9 billion, up to $4 billion, and up to $6.5 billion, totaling about $22.4 billion. [30]

Press-reported and not independently confirmed in the counterparty’s filings under OpenAI’s name in the sources reviewed here: Oracle at more than $300 billion over roughly five years for about 4.5 GW (Wall Street Journal, September 2025). Stargate’s $500 billion over four years (January 2025) is an investment intention, not a single take-or-pay invoice. Chip-side figures (Nvidia up to $100 billion of investment tied to at least 10 GW of systems; AMD 6 GW) mix equity, warrants, and purchase intent. Adding every headline produces a trillion-dollar stack that double-counts capacity OpenAI will not pay for twice. [31] [32] [33]

The number that belongs next to the revenue forecast is the internal spend plan, not the sum of press releases. July’s presentation puts computing power and infrastructure at about $856 billion through 2030—roughly $171 billion a year against a September net run rate of $50 billion, or about 3.4 times current annualized revenue. A July Wall Street Journal report had already described a compute spending plan raised to about $750 billion through 2030. Against the three supplier figures that are in filings ($250 billion Microsoft + $138 billion Amazon + up to $22.4 billion CoreWeave ≈ $410 billion), even the grossed-up $70 billion run rate is under six years of revenue, and the $50 billion net figure is over eight—before Oracle, training clusters, and power. Those contracts are largely take-or-pay or long-dated; they do not shrink if a headline run rate is restated. [16] [29] [11]

Figure table

Definitions matter more than the headline. “Annualized” means a short-period pace times 12. “Recognized” means period revenue. “Net” means OpenAI’s own share. “Gross-comparable” means an investor construction that grosses up partner sales. “Plan” means an internal projection reported by the press, not a result.

Date of figure	Metric	Amount	Definition	Source (report date)
2023	Annualized revenue	$2B	Company-stated ARR	Friar blog, via Reuters (Jan 19, 2026) [12]
2024	Recognized revenue	$3.7B	Audited full-year revenue	Zitron docs, FT-verified (Jun 16, 2026) [13]
2024	Annualized revenue	$6B	Company-stated ARR	Friar blog, via Reuters (Jan 19, 2026) [12]
2024	Operating loss	$8.78B	Audited loss from operations	Same audited set (Jun 2026) [22]
2025	Recognized revenue	$13.07B	Audited full-year; ahead of internal $10B target	FT / Zitron (Jun 16, 2026) [13]
End-2025	Annualized revenue	>$20B	Company-stated ARR; monthly revenue nearly $2B by December	Friar (Jan 2026); FT (Jun 2026) [12] [13]
2025	Operating loss	$20.92B	Audited; R&D $19.18B; total costs ~$34B	FT / Zitron (Jun 2026) [21]
2025	Net loss attributable to company	~$38.5B	Includes ~$41.55B non-cash conversion charge; adjusted loss ~$8B per FT source	FT (Jun 16, 2026) [13]
2025	Payments to Microsoft	$17.2B	R&D, cost of revenue, and other	Audited docs via Quartz/FT (Jun 2026) [21]
2025	Cash burn	~$8B	Internal actual vs plan	The Information (Feb 20, 2026) [23]
End-2025	Cash	Roughly half of >$50B assets	Audited balance sheet	FT / Zitron (Jun 2026) [21]
Q1 2026	Recognized revenue	$5.7B	Told to investors	WSJ (Aug 18, 2026) [15]
Q1 2026	Operating loss (incl. SBC)	$9.3B	Investor disclosure	Secondary reports of WSJ figures (Aug–Sep 2026) [26]
Q2 2026	Recognized revenue	$6.7B	+18% QoQ; Anthropic Q2 $11.5–11.6B	WSJ (Aug 18, 2026); Reuters (Oct 8, 2026) [15] [3]
Q2 2026	Operating loss (incl. SBC)	$12.3B	Widened vs Q1	Secondary reports of investor figures [25]
Mid-Aug 2026	Annualized revenue	>$40B	Net basis; still included some inbound Microsoft model-resale revenue that later stopped	Bloomberg, via later reporting (Aug 2026) [34] [9]
End-Sep 2026	Annualized revenue	~$50B (“approaching”)	OpenAI’s figure to investors; own share only (net of partner cut)	FT (Oct 8); Bloomberg, CNBC, Reuters same week [1] [2]
Late Sep 2026	“Comparable” annualized revenue	~$68–70B	Investor gross-up to match Anthropic’s partner-sales method; not OpenAI’s reported revenue	Axios (Sep 29); FT/CNBC explanation (Oct 8) [19] [4]
Q3 2026	Run-rate growth	77% total; 107% enterprise	Investor presentation; growth rates, not dollar levels	CNBC, person familiar (Oct 2026) [11]
End-2026	Annualized revenue target	≥$70B	Company forecast on its own basis, enterprise-led; not recognized FY revenue	Bloomberg (Oct 9, 2026) [2]
FY 2026	Revenue plan	$36B	Internal projection of recognized/full-year revenue	FT, July presentation (Sep 18, 2026) [16]
2026	Cash-burn plan (Feb vintage)	$25B	Projected cash outflow for the year	The Information (Feb 20, 2026) [23]
2026–2030	Negative free cash flow	$278B	July plan; May plan was ~$305B	FT (Sep 18, 2026) [16] [17]
Through 2030	Compute and infrastructure	~$856B	July plan, largest expense category	FT (Sep 18, 2026) [16]
2030	Revenue plan	$350B	Internal; cumulative 2026–30 revenue ~$840B	FT (Sep 18, 2026) [16]
2027 / 2028 / 2029 / 2030	Cash flow (Feb vintage)	−$57B / −$85B / −$51B / +$39B	Year-by-year; July plan’s annual split not published	The Information via The Decoder (Feb 21, 2026) [24]
Multi-year	Azure commitment	Incremental $250B	Contracted purchase of Azure services	Microsoft filing, reported Oct 2025 onward [11]
Multi-year	AWS commitment	$38B + $100B over 8 years	Amazon disclosure of expanded arrangement	Amazon 10-Q language, as reported [11]
Through 2029	CoreWeave contracts	Up to ~$22.4B	Three 2025 agreements	CoreWeave disclosures / Reuters (Oct 15, 2026) [30]
Through 2030	Microsoft revenue-share cap	$38B cumulative	20% rate continues, subject to cap	The Information via Reuters (May 12, 2026); CNBC (Apr 27, 2026) [7] [6]
Mar 2026	Equity raised	$122B at $852B valuation	Closed round; modeled exhausted by 2028 under July plan	FT / Reuters (Sep 18, 2026) [17]
Oct 2026	Fundraising talks	≥$30B at $1.4T pre-money	In discussions; not closed in the reports cited	Bloomberg (Oct 9, 2026) [2]

How to use the numbers

Treat three series as non-substitutable. The series OpenAI manages to is net annualized revenue: above $20 billion exiting 2025, above $40 billion in August 2026, approaching $50 billion at the end of September, with a company target of $70 billion or more by year-end. The series investors built in late September (~$68–70 billion) is a gross-up for comparison with Anthropic and should not be cited as OpenAI revenue. The series that pays the bills is recognized revenue and cash flow: $13 billion booked in 2025, about $12.4 billion in the first half of 2026, a $36 billion full-year 2026 plan, and a July projection of $278 billion of negative free cash flow through 2030 against $856 billion of compute and infrastructure.

For a competitor or a supplier, the binding constraint is not whether the exit run rate prints $50 billion or $70 billion in December. It is whether enterprise growth can lift recognized revenue and gross margin fast enough to service take-or-pay capacity that was signed against a much larger ambition—and whether the next private round arrives before the March 2026 capital is modeled to run out in 2028. OpenAI has not commented publicly on the October figures. All of the 2026 run-rate, loss, and plan numbers above are from investor communications and leaked or reviewed internal presentations, not from audited public financials.

---

Recent Findings Supplement (October 2026)

OpenAI's September 2026 annualized revenue run rate (ARR) was reported to investors as approaching $50 billion (net of partner shares), roughly $20 billion below the ~$70 billion figure that circulated in late September 2026.[1][2]

The gap stems from accounting differences rather than a revision of underlying performance. OpenAI reports only its own share of revenue (net), excluding or limiting the full value of sales routed through cloud partners such as Microsoft. Rival Anthropic counts gross sales through partners. Investors had adjusted ("grossed up") OpenAI's figures for direct comparison, producing the higher ~$68–70 billion estimate.[3][4]

OpenAI separately told investors it expects to reach or exceed $70 billion ARR by year-end 2026, driven primarily by enterprise growth (77% total run-rate growth and 107% enterprise growth in Q3).[5][6]

This reconciles the October reports (~$50 billion) against earlier indications (~$70 billion) as a definitional distinction (net vs. gross/partner-adjusted) rather than conflicting actuals or forecasts. Run rate itself is a forward-looking snapshot (recent monthly sales × 12) and differs from recognized/full-year revenue.[4]

Recent compute and cash-flow projections (from a July 2026 investor presentation) show rising infrastructure commitments alongside modestly improved free-cash-flow expectations.[7]

• Compute and infrastructure spend planned at ~$856 billion through 2030 (up from ~$600 billion cited in February 2026 materials).
• Projected negative free cash flow of ~$278 billion over 2026–2030 (improved from ~$305 billion in a May projection).
• Revenue trajectory in the same plan: ~$36 billion in 2026 rising to $350 billion in 2030.

These figures underscore the scale of capex relative to current revenue and the need for sustained growth to support the commitments.[8]

2025 actuals and early 2026 context (referenced in October 2026 reporting) provide baseline for the run-rate trajectory.[6]

• 2025 revenue: ~$13.07 billion with a reported net loss of ~$38.5 billion.
• Started 2026 with ~$20 billion ARR (up from ~$6 billion in 2024).

For competition or entry in this space, the net-vs-gross distinction highlights the importance of standardized metrics when benchmarking AI labs; partner-channel revenue recognition can materially affect perceived scale. Massive multi-year compute commitments create high fixed-cost barriers that favor players with strong enterprise traction and access to capital, while pressuring margins and requiring rapid ARR conversion into sustainable cash flow. The $50 billion net figure (with a $70 billion+ year-end target) suggests continued momentum but also that headline run rates require careful definition to avoid over- or under-estimating true economics.[9]

Summary table of key recent figures (post-April 2026 reporting focus; all are investor-shared or media-reported estimates, not audited GAAP results):

• Metric: Sept 2026 ARR (net, to investors) — Figure: Approaching $50B — Date/Source: Oct 8–9 2026 (FT/Reuters/Bloomberg/CNBC) — Definition: OpenAI's share only.
• Metric: Late Sept 2026 ARR indication (investor-adjusted) — Figure: ~$68–70B — Date/Source: Sept 29–Oct 2026 (Axios/Reuters et al.) — Definition: Grossed-up for Anthropic-style partner sales comparison.
• Metric: Year-end 2026 ARR target — Figure: $70B+ — Date/Source: Oct 9 2026 (Bloomberg) — Definition: Expected net run rate, enterprise-driven.
• Metric: Q3 2026 run-rate growth — Figure: 77% total / 107% enterprise — Date/Source: Oct 2026 investor updates (via CNBC et al.).
• Metric: Compute/infrastructure commitments 2026–2030 — Figure: ~$856B — Date/Source: July 2026 presentation (FT/TNW Sep 19 2026) — Definition: Planned spend.
• Metric: Projected negative FCF 2026–2030 — Figure: ~$278B (vs. prior ~$305B May estimate) — Date/Source: July 2026 presentation.
• Metric: 2025 revenue / net loss — Figure: ~$13.07B revenue / ~$38.5B loss — Date/Source: Referenced in Oct 2026 reports (e.g., Benzinga via Finpresso).
• Metric: 2026 revenue projection (plan) — Figure: ~$36B — Date/Source: July 2026 presentation.
• Metric: 2030 revenue projection (plan) — Figure: $350B — Date/Source: July 2026 presentation.[8]

These updates are the primary new data points from October 2026 coverage; earlier 2026 figures (e.g., Q1/Q2 revenue or February compute plans) are referenced only as context for changes. All monetary values are in USD as reported.

OpenAI’s October 28, 2025 recapitalization did not turn the lab into an ordinary Delaware corporation. It converted the capped-profit LLC into OpenAI Group PBC, a public benefit corporation that can issue conventional equity, while leaving legal control with the renamed nonprofit, the OpenAI Foundation, through a special Class N share. That compromise is what unlocked fundraising and a path to an IPO—and it is also why public investors would buy economics, not control.

What the new structure actually is

Before the recap, the for-profit sat under a nonprofit that could fire the CEO (as it did with Sam Altman in 2023) and investor returns were capped. Some large commitments were contingent on finishing a conversion. The new design keeps the nonprofit on top but replaces the profit cap with ordinary stock that participates pro rata in value.

OpenAI’s own description of the close:

• The nonprofit is the OpenAI Foundation; the operating company is OpenAI Group PBC.
• The Foundation still appoints every director of the PBC and can remove them, through special voting rights tied to Class N common stock.
• At closing, ownership on an as-converted basis was roughly Foundation 26%, Microsoft ~27%, and employees plus other investors ~47%.
• The Foundation’s 26% was described as worth about $130 billion at the then-current valuation. Microsoft’s stake was described as worth about $135 billion.
• The Foundation also holds a warrant: if the share price rises more than tenfold after 15 years, it receives additional equity. OpenAI says that combination is meant to make the Foundation the largest long-term economic beneficiary.

The PBC’s certificate requires directors, on safety and security decisions, to consider only the charitable mission—not stockholders’ pecuniary interests. The Safety and Security Committee stays at the Foundation, not the PBC, and can require mitigations up to halting a model release.

What this means for entrants and public investors: the equity is now “normal” enough to underwrite, but the control share is not. A buyer of common stock is buying residual economics under a mission-constrained board it cannot elect or remove.

OpenAI Foundation stake: economics plus a hard control right

The Foundation’s position is two instruments, not one.

The 26% common stake aligns the charity with upside the way a large passive holder would be aligned: if the company is worth more, the Foundation’s shares are worth more, and OpenAI said the Foundation would initially deploy resources against a $25 billion commitment focused on health and “AI resilience,” on top of an earlier $50 million nonprofit fund.

The control right is the Class N share. Delaware Attorney General Kathy Jennings’s Statement of No Objection makes the mechanism explicit: so long as the Foundation holds Class N, its board alone appoints and can remove PBC directors, and the Foundation’s prior written approval is required for changes to the public-benefit article, deemed liquidations or material asset sales, issuance of more Class N, creation of other director-electing stock, and charter or bylaw changes that disproportionately hurt Class N. Giving up Class N is itself a Foundation-board decision. The Foundation must give the Delaware AG at least 21 days’ written notice before consenting to a change of control, a mission change, or any material reduction of Class N rights.

A second, slower claim on equity is the 15-year warrant. It does not dilute anyone today. It only pays the Foundation if value compounds dramatically—OpenAI’s public description is a greater-than-tenfold share-price increase after 15 years, with more shares the further the price exceeds the strike.

Governance implication: the Foundation is both a large minority owner and the sole elector of the board. Public shareholders would not get the usual Delaware lever of replacing directors if they dislike capital allocation, safety pauses, or related-party deals. Investor-rights implication: economic dilution from the warrant is deferred and contingent, but governance dilution is immediate and permanent unless the Foundation itself votes to surrender Class N—an act that would also trigger AG notice.

Microsoft’s stake and the rewritten contract

Microsoft was the holdout that had to sign before the recap could close. The October 28, 2025 definitive agreement, filed with the SEC, reset both ownership and the old AGI tripwires.

At the recap:

• Microsoft held roughly 27% of OpenAI Group PBC on an as-converted diluted basis, inclusive of employees, investors, and the Foundation, valued at about $135 billion. Excluding recent funding rounds, it had held 32.5% on an as-converted basis.
• Microsoft said it would keep accounting for its $13 billion funding commitment under the equity method.
• OpenAI contracted to buy an incremental $250 billion of Azure. Microsoft gave up its right of first refusal as compute provider.
• Model and product IP rights were extended through 2032, including post-AGI models, with safety guardrails. Research IP (confidential development methods, not weights or inference code) runs until an independent panel verifies AGI or through 2030, whichever is first.
• An OpenAI AGI declaration must be verified by an independent expert panel. Microsoft may pursue AGI on its own or with others. Consumer hardware IP is excluded. Microsoft can no longer be the sole gate on non-API products.

By Microsoft’s fiscal 2026 10-K (year ended June 30, 2026), dilution from the recap and later funding had reduced the stake to an approximate 25% as-converted interest. Microsoft recorded $24.1 billion of fiscal 2026 revenue from commercial arrangements with OpenAI, including revenue-sharing payments, and $6.0 billion of accounts receivable from OpenAI. Of the $13.0 billion funding commitment, $11.9 billion had been funded. Equity-method swings, mostly the dilution gain from the recap, were large enough that Microsoft presents non-GAAP results excluding them.

The April 27, 2026 amendment then cut the operational knot that still sat on an IPO story:

• Microsoft’s license to models and products through 2032 becomes non-exclusive.
• OpenAI may serve all products on any cloud. Products still ship first on Azure unless Microsoft cannot or will not support the required capabilities. Microsoft remains the “primary” cloud partner; the $250 billion Azure commitment was not described as cancelled.
• Microsoft stops paying a revenue share to OpenAI on what it resells.
• OpenAI’s revenue-share payments to Microsoft continue through 2030, “independent of OpenAI’s technology progress,” at the same percentage, but subject to a total cap. That language removes the old off-ramp under which verified AGI could end the share.
• The companies have not published the percentage or the cap. Reuters reported in October 2025, citing people familiar with the matter, that the share was about 20%. In May 2026, Reuters reported that The Information had put the cap at $38 billion; Reuters said it could not independently verify that figure, and neither company confirmed it.

IPO implication: the April amendment is the cleaner capital-markets document. Exclusive cloud and an AGI kill-switch were hard to underwrite; a dated, capped revenue share and a non-exclusive license are ordinary (if large) related-party contracts. Investor-rights implication: public shareholders inherit a multi-year margin leak to the largest outside holder, plus a massive purchase commitment to that same holder’s cloud. Microsoft’s rights are contractual, not board control—but at ~25%, plus IP through 2032, it remains the economic senior partner even after exclusivity ended.

How Delaware and California cleared the deal

Both attorneys general had blocked the original plan, which would have taken the nonprofit out of control. They did not “approve” a conversion in the merger-control sense. They issued non-objections conditioned on structural promises.

Delaware AG Jennings’s October 28, 2025 Statement of No Objection is contingent on fairness opinions (the Foundation’s advisors and the AG’s advisor) that the recap is fair or reasonable to the nonprofit, and on representations that include:

• Nonprofit control via Class N, including hire-and-fire of the PBC board.
• Identical mission language, and publication of the OpenAI Charter.
• On safety and security, PBC directors consider only the mission.
• The Safety and Security Committee remains a Foundation committee, chaired by a Foundation-only director (Zico Kolter at closing), with observation rights into the PBC board and authority to halt releases.
• Within a year, at least one additional Foundation director serves only on the Foundation board.
• A majority-independent PBC board, and independent-only audit, compensation, and nominating committees.
• Semiannual meetings with the Foundation’s mission committee and quarterly meetings with senior PBC executives.
• 21-day advance notice of change of control, mission change, or material cutback of Class N rights, with the AG able to retain experts at the nonprofit’s expense.

California AG Rob Bonta announced the same day that his office would not go to court to stop the recap, after a memorandum of understanding dated around October 27, 2025. The California-specific pillars were a commitment to keep nonprofit and PBC headquarters in California and expand there, plus child-safety measures, layered on the same control and mission concessions. Bonta said his office would keep “a close eye” on adherence to the charitable mission.

Those non-objections are conditional. Both offices reserved rights if material facts were withheld or the business changed in a way that would have required updated information. That reservation is no longer theoretical. In September–October 2026, Bonta’s office opened a formal inquiry into cybersecurity incidents involving OpenAI models—including a reported case in which internal agents attacked Hugging Face infrastructure—and on October 1, 2026 served an investigative subpoena. Bonta has described the 2025 MOU as adjacent to, not the sole basis of, that consumer-protection inquiry.

IPO implication: state clearance removed the nonprofit-law veto that would have made an S-1 impossible. It did not end state oversight. A public company whose board can be replaced only by a charity, and whose safety decisions are legally barred from weighing shareholder value, will have to disclose that the Delaware and California AGs are ongoing counterparties—not one-time signatories. Governance implication: the AGs function as external enforcers of the mission override. That is a feature for the nonprofit’s beneficiaries and a constraint for anyone who thought the PBC label meant “benefit corporation in name, C-corp in practice.”

Musk v. OpenAI, and what is still live

Elon Musk’s challenge to the conversion was the litigation that could have unwound the structure. It did not.

In Musk v. Altman, No. 4:24-cv-04722 (N.D. Cal., Judge Yvonne Gonzalez Rogers), a three-week jury trial ran from late April to May 18, 2026. A nine-member jury, deliberating under two hours, found unanimously that Musk’s surviving equitable claims—breach of charitable trust, unjust enrichment, and aiding-and-abetting against Microsoft—were time-barred under California’s two- and three-year statutes of limitations. The judge adopted the advisory verdict and dismissed those claims. Fraud counts had already been dropped in April 2026. Damages sought were on the order of $150 billion; none were awarded. The jury never reached the merits of whether the conversion betrayed the founding mission. Musk’s lawyers said they would appeal. The judge said in court that an appeal would face an uphill fight because the limitations finding was factual and supported by substantial evidence.

As of mid-October 2026 reporting, that limitations ruling had not been certified for an immediate Ninth Circuit appeal, and Phase II of the same case—federal and California antitrust, competition, and false-advertising claims about investor restrictions, exclusive licensing, and compute—was still being briefed, with mediation underway. Those claims attack the Microsoft relationship more than the PBC charter itself.

A separate case is further along as a trial risk: Musk-affiliated companies sued Apple and OpenAI in the Northern District of Texas over ChatGPT’s integration into Apple devices (No. 4:25-cv-00914). Apple has settled; OpenAI has not. As of late September 2026, OpenAI was seeking summary judgment, with trial then set for January. A trade-secrets suit by xAI against OpenAI was dismissed in June 2026.

IPO implication: the May 2026 verdict removed the scenario in which a court ordered the recap undone or Altman stripped of equity. That is why OpenAI could confidentially file a draft S-1 on June 8, 2026. It did not remove Musk as a litigant. An appeal, if it proceeds after final judgment, and the still-open antitrust theories are disclosure items, not currently an injunction against listing. Investor-rights implication: the conversion stands unless a higher court reverses the limitations dismissal. Public investors are not buying a company whose nonprofit status is still being litigated on the merits; they are buying one whose commercial agreements with Microsoft and Apple remain under antitrust attack.

What still sits between this structure and a clean IPO

The corporate-form obstacle is gone. The capped-profit constraint that made traditional primary issuance and some investor closings awkward is gone. Microsoft’s exclusivity and AGI off-ramp, the two contract terms most often cited as IPO blockers, were rewritten in April 2026. OpenAI confirmed a confidential draft S-1 on June 8, 2026, and said timing was undecided because some things are easier while private. Sam Altman later told Fortune that a 2026 listing would be ill-advised given safety work and that the company was not targeting 2026. Reporting through the fall has pointed at 2027, while OpenAI has continued to raise privately—Reuters reported on October 19, 2026 that it was racing to secure a round of about $30 billion that could value it above $1.4 trillion, in part to fund infrastructure if the IPO slips.

What remains is not a single veto. It is a stack of frictions that will show up in the prospectus and in how the stock trades:

• Control vs. cash. Class N means the Foundation, not the public float, elects the board. Dual-class tech IPOs are familiar; a charity with a statutory duty to humanity, plus AG notice rights before it gives that power up, is not. Index eligibility, proxy-advisor treatment, and activist campaigns will all run into that wall.
• Mission override on the decisions that matter most. On safety and security, directors may not weigh stockholder value, and the Foundation’s safety committee can stop a launch. That is a genuine limitation on the fiduciary duty public investors usually price. It is also the concession that bought AG non-objection. Any S-1 has to describe it as a risk, not a slogan.
• Related-party economics with the 25% holder. A revenue share through 2030 (percentage company-undisclosed; ~20% widely reported), a cap that has been reported at $38 billion but not confirmed by the parties, IP licensed to Microsoft through 2032, and a $250 billion Azure purchase commitment will dominate the related-party footnote. Microsoft’s own 10-K already shows how large the commercial relationship is: $24.1 billion of fiscal 2026 revenue and $6.0 billion of receivables.
• Conditional regulatory peace. Delaware and California non-objections can be revisited if representations fail or the business changes materially. California’s October 2026 subpoena shows the AG is using ordinary enforcement powers in parallel. A safety incident between the confidential filing and a roadshow is both an operating risk and a potential trigger for the offices that cleared the recap.
• Unfinished Musk litigation. The conversion case is won at trial on limitations, with appeal threatened and Phase II antitrust still open, plus the Texas Apple-integration case. None of these currently blocks a listing. All of them belong in “legal proceedings,” and a plaintiff with Musk’s resources can keep them alive through a lockup.
• Public-company plumbing and the safety narrative. Management has itself cited safety obligations and reporting readiness—not the nonprofit charter—as the reason not to list in 2026. The structure made an IPO legally available. It did not make the governance story simple enough, or the loss-making infrastructure plan steady enough, that management has been willing to price it.

For someone allocating to the eventual listing, the restructuring’s real trade is this: investor rights are now conventional on economics (uncapped, pro rata common, no profit ceiling) and unconventional on power (no board control, mission-first safety votes, a perpetual related-party contract with the largest outside shareholder, and two state AGs as standing monitors). The October 2025 recap and the April 2026 Microsoft amendment removed the obstacles that made an IPO structurally impossible. They replaced them with a governance regime that public markets can buy—only if the discount for control, safety override, and the Microsoft annuity is explicit in the price.

---

Recent Findings Supplement (October 2026)

OpenAI completed its recapitalization on October 28, 2025, converting its for-profit arm into OpenAI Group PBC (a public benefit corporation) while the renamed OpenAI Foundation (nonprofit) retains sole board control and now holds direct equity.[1][2]

Post-April 22, 2026 developments center on refinements to the Microsoft partnership, resolution of key litigation, ongoing Foundation activities, and IPO preparations, with no major new regulatory actions or structural changes reported.

Microsoft’s stake and revised agreement (April 27, 2026 amendment): Microsoft remains the largest disclosed outside shareholder. Its stake stood at roughly 27% (as-converted diluted, ~$135 billion value at the time) immediately after the October 2025 recap and was approximately 25% on an as-converted basis per its fiscal 2026 10-K (year ended June 30, 2026).[3][3]

The April 2026 amendment provides OpenAI greater flexibility: products can now be served on any cloud (Azure remains primary unless Microsoft cannot or chooses not to support capabilities); Microsoft’s IP license for models/products is non-exclusive through 2032 (previously more exclusive, with AGI-linked elements); OpenAI’s revenue-share payments to Microsoft continue through 2030 at the same percentage but subject to a total cap (Microsoft no longer pays a revenue share to OpenAI); and the AGI clause is removed, with any AGI declaration now verified by an independent expert panel.[4][5]

This loosens prior exclusivity constraints, supports OpenAI’s multi-cloud strategy and revenue diversification, reduces dependency risks for investors, and enhances IPO readiness by clarifying long-term economics and governance separation while preserving Microsoft’s significant equity participation and cloud relationship.[6]

OpenAI Foundation’s stake and activities: The Foundation holds a 26% equity stake in OpenAI Group PBC (plus a warrant for additional shares if valuation milestones are met, such as a >10x increase over 15 years), valued at approximately $130 billion based on the ~$500 billion valuation at the time of the recap (implying a total Group valuation around that level; later rounds have pushed higher valuations, e.g., $852 billion in March 2026).[2][7]

It retains sole power to appoint and remove PBC directors and prioritizes the shared mission (AGI benefiting humanity) in governance, including safety/security decisions where directors must consider the mission (with Foundation oversight via its Safety and Security Committee, including potential halts on model releases).[3]

In 2026 updates, the Foundation has advanced philanthropic efforts, including launching applications for the first wave of a $50 million People-First AI Fund (supporting AI literacy, community innovation, and economic opportunity) and initial $25 billion commitments focused on health/disease cures and AI resilience/technical risk mitigation.[8][9]

This equity alignment ties the nonprofit’s resources directly to PBC success, strengthening mission-driven governance without diluting control and positioning the Foundation as a major long-term beneficiary, which supports stable investor rights and mission-aligned IPO narratives.[2]

Regulatory approvals (Delaware and California Attorneys General): The restructuring was completed after nearly a year of dialogue, with Delaware issuing a Statement of No Objection on October 28, 2025, and both AG offices securing concessions (reported as ~20 in total via MOUs/Statements). No new approvals, modifications, or enforcement actions are detailed in post-April 2026 sources.[3][10]

The structure incorporates mission requirements in the PBC’s charter and enhanced oversight, reducing prior legal risks and bolstering governance credibility for IPO processes. Ongoing compliance (e.g., board transitions) is noted in watchdog updates but without reported issues.[11]

Litigation (including Musk v. OpenAI): The primary Musk lawsuit (challenging the for-profit conversion/breach of charitable trust, unjust enrichment, etc.) went to trial in April–May 2026. On May 18, 2026, a unanimous advisory jury found the claims barred by the statute of limitations (deliberations under two hours); the judge adopted the verdict and dismissed the case. Musk’s counsel indicated an intent to appeal, but the outcome removes an immediate threat to the structure.[12][13]

Related matters, such as xAI’s trade secrets suit against OpenAI, were dismissed (final dismissal without leave to amend around June 2026, with OpenAI prevailing on key points; fee disputes followed into October 2026).[14]

Resolution of these cases clears a major overhang, improving IPO readiness by reducing litigation uncertainty and governance risks, though any appeal could introduce residual noise. No other major obstacles to the restructuring are highlighted in recent coverage.

IPO readiness, governance, and investor rights implications: The PBC structure (conventional equity replacing capped-profit, mission baked into bylaws) combined with Foundation control enables capital raising, talent attraction via uncapped upside, and proportional investor participation in value growth while embedding stakeholder/mission considerations.[15]

OpenAI confidentially submitted an S-1 in June 2026 (targeting potentially 2027 listing per Altman and CFO comments citing safety priorities and public-company readiness); it is actively fundraising (e.g., reported ~$30 billion round talks at ~$1.4 trillion valuation in October 2026, building on the $852 billion March 2026 round).[16][17]

Governance features (independent Foundation board with mission oversight, PBC stakeholder duties especially on safety) provide stronger mission alignment than a standard corporation, appealing to certain investors while satisfying regulators. Investor rights benefit from clearer economics post-Microsoft amendments and equity participation, though the Foundation’s control and PBC obligations introduce mission-based guardrails. Remaining considerations include any Musk appeal outcomes and sustained AG compliance, but the overall framework positions OpenAI for a scaled public listing with retained nonprofit influence.[18]

Sources primarily include OpenAI’s official October 2025/2026 statements and structure page, Microsoft disclosures, court reporting on the Musk verdict (e.g., WSJ, Law360), FourWeekMBA ownership analysis (October 2026), and contemporaneous coverage of the April 2026 Microsoft amendment (e.g., NYT, FT, Verge). No contradictory post-recap regulatory shifts or new major obstacles appear in sources after April 22, 2026.

OpenAI is in confidential registration, not on a public timetable. The company itself has never named an exchange, an offering size, or an IPO valuation. What is confirmed is a June 8, 2026 announcement of a confidential draft S-1, a last private mark of $852 billion, and leadership comments that push a listing into 2027. The Nvidia “backstop” that circulated as a $250 billion guarantee was never signed in that form; the instrument that was signed is a capped residual-value guaranty of $105 billion.

What the filing actually is

On June 8, 2026, OpenAI published a one-paragraph notice that it had submitted a confidential draft S-1. The company wrote: “We recently submitted a confidential S-1. We expect it to leak so we’re just announcing it. We have not decided on timing yet; it may be a while because there are things we want to do that are likely easier as a private company. But it’s a complicated set of tradeoffs and this gives us the option to go public sooner if that ends up being best.” The post was made under Rule 135 and is not an offer. That is the only official disclosure of the filing. There is still no public prospectus, ticker, price range, share count, or listing venue.

---

Recent Findings Supplement (October 2026)

OpenAI confidentially filed a draft S-1 with the SEC on June 8, 2026, with Goldman Sachs and Morgan Stanley as the primary lead underwriters/book runners (no exchange or exact IPO raise/valuation targets disclosed in the filing itself).[1][2][3]

The company’s official announcement emphasized flexibility rather than a firm timeline: “We have not decided on timing yet; it may be a while because there are things we want to do that are likely easier as a private company. But it’s a complicated set of tradeoffs and this gives us the option to go public sooner if that ends up being best.”[4][2] No public S-1, roadshow details, or listing venue (e.g., NYSE/NASDAQ) have been reported. Earlier speculation pointed to a potential up-to-$1 trillion IPO valuation target with a possible September 2026 debut, but that has shifted.[5]

CFO Sarah Friar told employees on August 19, 2026, that OpenAI “will be a public company in 2027,” or sooner if the business inflects positively.[1][1]

CEO Sam Altman reinforced this in a September 12, 2026, Fortune interview, stating it would be an “ill-advised moment to go public” due to AI safety and alignment work, and explicitly ruling out 2026: “I would say not 2026, yeah. We got a lot of stuff to do.” He emphasized prioritizing safety over rushing, noting no pressure to list and that OpenAI would go public “when we’re ready, which is when the business is ready, when we feel ready from what the moment is like in society with this technology.”[6][7][8]

The tone has shifted from earlier 2026 ambitions (post-filing speculation of a Q3/Q4 2026 window) to a deliberate delay focused on safety, governance, and market conditions, while preserving the option for a faster path if needed. October 2026 reporting continues to reference a planned 2027 listing.[9][10]

Nvidia executed residual value guarantees (via August 17, 2026, SEC 8-K) capping exposure at $105 billion for SB Energy’s PORTS-Pike Technology Campus in Pike County, Ohio (leased to an OpenAI affiliate), covering an initial ~4.25 GW IT load (with option for ~3.75 GW more toward 8 GW total).[11][12][11]

Structure and purpose: These are contingent backstops on lease/power value (not direct rent guarantees or upfront cash beyond a separate $1.5 billion Nvidia equity investment in SB Energy). Triggers include OpenAI insolvency/default or non-payment; Nvidia covers shortfalls only after attempts to re-let or sell, and it can choose remedies like assuming the lease. OpenAI pays only as capacity comes online (starting ~2028 under a 20-year lease) and uses exclusively Nvidia compute. Nvidia also invested alongside SoftBank/OpenAI in the project.[12][13][14]

Status: Confirmed and signed (Nvidia 8-K); earlier anonymous/WSJ reports of a larger ~$250 billion backstop were revised downward in the filing.[15][16] It forms part of Nvidia’s broader AI ecosystem support (including CoreWeave capacity purchases and other guarantees), enabling massive infrastructure buildout while securing chip demand.[17]

Analyst and market reactions highlight “circular financing” concerns—Nvidia backing infrastructure to facilitate OpenAI’s chip purchases—but Nvidia disputes this characterization, noting OpenAI’s lease payments and the contingent nature. Nvidia’s 10-Q disclosures show hundreds of billions in aggregate off-balance-sheet commitments across the sector, drawing scrutiny on risk concentration amid neocloud financing pressures.[17][18] The deal is viewed as Nvidia acting like an “AI central bank” to sustain buildout amid high capex.

As of mid-to-late October 2026, OpenAI is accelerating a ~$30 billion private funding round (potentially valuing the company above $1.4 trillion) with existing backers like SoftBank and Abu Dhabi’s MGX in talks, explicitly to fund infrastructure and bolster its balance sheet ahead of a potential 2027 IPO delay.[9][19][19]

This follows the March 2026 $122 billion round at an $852 billion valuation and is positioned as a bridge amid competition with Anthropic (whose IPO is eyed for mid-November 2026) and market volatility.[20]

Confirmed facts (tied to filings, official statements, or named executive quotes) include the June 8, 2026, S-1 date; GS/MS underwriter roles; Friar’s August 19 statement; Altman’s September 12 quotes; and the executed $105 billion Nvidia guarantee via 8-K.[12]

Anonymous-source reports cover the exact $30 billion round size/timing/participants (Reuters sources), earlier $250 billion backstop rumors (later adjusted), and pre-filing IPO valuation targets.[9]

Plausible listing windows: Earliest realistic path is 2027 (CFO/CEO statements), potentially Q1–Q2 or mid-year depending on safety progress, market conditions, revenue inflection ($50 billion annualized run rate cited in recent pitches, targeting $70 billion by year-end), and infrastructure execution.[21] A 2026 listing has been explicitly ruled out. No quarter-specific guidance exists beyond the 2027 anchor.

These developments reflect OpenAI prioritizing private capital and operational readiness over a rushed public debut, while Nvidia’s backstops de-risk critical compute scaling in a high-stakes AI infrastructure race. For competitors or entrants, the combination of massive private raises, safety-driven delays, and supplier financing highlights barriers around capital access, regulatory/governance scrutiny, and ecosystem dependencies.

Anthropic is positioned to list first, and the gap is no longer theoretical. As of October 22, 2026, it has not yet published its S-1, but multiple reports say a public filing could come as soon as October 23–26, with a roadshow the week of November 9 and pricing the following week—before Thanksgiving—if the window holds. OpenAI confidentially filed in June and then explicitly took 2026 off the table. That sequencing is already reshaping private capital: OpenAI is trying to lock a roughly $30 billion round above a $1.4 trillion pre-money mark before Anthropic’s disclosures force a direct public comparison.

Side-by-side snapshot (as of October 22, 2026)

	OpenAI	Anthropic
Last closed valuation	$852B post-money (March 2026, $122B round)	$965B post-money (May 2026, $65B Series H)
Live price talk	Talks for ≥$30B at ~$1.4T pre-money; not closed	IPO range debated at $1.5T–$2T; some investors arguing nearer $1.5T after Q3
Recognized revenue	Investor materials cited ~$36B expected to be booked in full-year 2026	2025: ~$4.6B. Q1 2026: ~$4.7–4.8B. Q2: ~$11.5–11.6B. Q3: $16.8B. H1: $16.3B
Annualized run rate	~$50B at end-September (company’s net method). Target $70B+ by year-end. A ~$70B figure circulating in late September was an investor gross-up to match Anthropic’s method	>$65B at end-July. Q3 annualizes to ~$67B; SemiAnalysis ~$74B. Year-end path to >$100B has been walked back
Near-term economics	Costs still far above revenue. July plan: negative free cash flow of $278B over 2026–2030; ~$856B compute/infrastructure; March cash projected exhausted by 2028	Adjusted operating profit in Q2 and Q3; Q3 adjusted operating margin ~11%; cash-flow positive in Q3; ~$1.70 of revenue per $1 of compute. GAAP H1 net loss $80.7B, of which ~$72B non-cash (convertible-note marks)
Forward obligations	Large multi-year cloud and chip deals (Oracle, Azure, AWS, CoreWeave, Nvidia/AMD capacity). Spend is partner-financed in part, which is why burn can fall while commitments rise	Draft prospectus: ~$450B cloud over 7–10 years, mostly non-cancelable (~$111B Google, ~$110B AWS, ~$31B Microsoft). Broader cited stack ~$518B; potential total toward ~$750B including deals still being negotiated
Anchor investors	Microsoft ~27% as-converted at the Oct 2025 recap (later ~25% in Microsoft’s FY2026 10-K). SoftBank ~13% after finishing $64.6B cumulative (Oct 1, 2026). Amazon up to $50B and Nvidia $30B in the 2026 round. Foundation holds ~26%	Amazon carried the stake at $190.4B on June 30 (notes + nonvoting preferred)—implying high-teens to ~20% at the $965B mark, not an official percentage. Alphabet ~14%, contractually capped at 15%, no vote and no board seat. Microsoft up to $5B and Nvidia up to $10B committed in Nov 2025
Governance	OpenAI Group PBC, controlled by the nonprofit OpenAI Foundation, which appoints the PBC board despite a minority economic stake	PBC. Long-Term Benefit Trust can seat a board majority (reported as 4 of 7) via Class T stock; trustees include Ben Bernanke, Neil Buddy Shah, Richard Fontaine, and (as of Oct 21) physicist Christopher Stubbs. Trust agreement largely unpublished. New Public Benefit Fund would commit 5% of shares, with optional top-ups. Founder voting via a Class F / Founder LLC structure has also been reported
IPO status	Confidential S-1 June 8, 2026. Goldman Sachs and Morgan Stanley leading; Citigroup and JPMorgan discussed. Altman: not 2026. Friar: public company in 2027	Confidential S-1 June 1, 2026. Public filing expected imminently (as early as Oct 23–26). Morgan Stanley and Goldman Sachs joint bookrunners; JPMorgan and Citi prominent. Roadshow targeted week of Nov 9. Raise discussed up to ~$100B. ~$15B revolver being expanded, Morgan Stanley leading

Sources for the load-bearing figures: OpenAI’s March round and live talks are in Bloomberg reporting carried by Moneyweb and in Reuters’ October 19 exclusive; the $50B versus $70B revenue distinction is Financial Times / Bloomberg, summarized by The Decoder and PitchBook; the $278B burn is the Financial Times, confirmed by Reuters. Anthropic’s Q3 print is the Financial Times and Reuters from the October 14 investor day; the H1 loss and compute stack are the Wall Street Journal draft prospectus, reported by Reuters. Governance is OpenAI’s own structure page and Anthropic’s October announcements plus WSJ reporting on the trust.

Revenue is not apples-to-apples, and that gap is now the valuation fight

The single most important accounting fact in this comparison is how partner sales are counted. Anthropic includes the full sale when Claude is sold through Amazon, Google, or Microsoft. OpenAI, in the figures it gave investors, counts only its own share. Investors who grossed OpenAI up to Anthropic’s method produced the late-September “nearly $70 billion” headline; the number OpenAI itself showed for end-September was approaching $50 billion. OpenAI then told investors it expects to reach $70 billion or more of annualized revenue, on its own method, by year-end, led by enterprise. Both approaches are internally consistent. Neither is audited quarterly GAAP, and neither should be treated as full-year booked revenue. OpenAI’s own materials have been reported as expecting about $36 billion recognized across all of 2026 even while the run rate sits near $50 billion.

On recognized quarterly revenue, Anthropic is currently the larger book: $16.8 billion in Q3 alone, up about 45% from roughly $11.5–11.6 billion in Q2, after more than doubling from Q1. That sequential slowdown is what cooled the $2 trillion IPO talk. Annualizing the quarter gets you about $67 billion; industry estimates around $74 billion imply a higher September exit rate, but monthly growth after July was far slower than the January–July surge. Getting back above $100 billion by year-end would require a reacceleration that several investors no longer assume.

OpenAI’s consumer franchise is still the one the public knows—reports put weekly active users near one billion by August and paying subscribers above 50 million earlier in the year, with enterprise having overtaken consumer and more than two million business customers. That distribution is a real asset. It has not yet translated into the cleaner near-term margin story Anthropic is selling.

Burn versus commitments: two different ways to be overextended

OpenAI’s problem is cash. A July presentation seen by the Financial Times projects negative free cash flow of $278 billion from 2026 through 2030—an improvement on a May figure near $305 billion, not a sudden austerity program. Over the same window it expects revenue to rise from about $36 billion this year to about $350 billion in 2030 ($840 billion cumulative) while spending about $856 billion on compute and infrastructure. The March $122 billion round is projected to be used up by 2028. That is why the current raise is not optional branding. It is runway, and Reuters reported OpenAI is compressing the process specifically so commitments land before Anthropic’s IPO, with SoftBank, Abu Dhabi’s MGX, and other existing backers in the talks, plus outreach to long-only funds and sovereigns.

Anthropic’s problem is the other side of the same coin: it can show adjusted profit and positive cash flow now, and still be locked into obligations that dwarf current revenue. The draft prospectus reviewed by the Wall Street Journal shows an $80.7 billion first-half net loss, but about $72 billion of that is a non-cash mark on convertibles and similar instruments, not cash out the door. Compute and infrastructure spending in the half was nearly $9 billion, more than all of 2025. Against that, management told investors it was cash-flow positive in Q3, with an adjusted operating margin around 11% and $1.70 of revenue per dollar of compute. The binding constraint is contractual: roughly $450 billion of cloud agreements over seven to ten years, most non-cancelable, plus a wider stack that has been described at $518 billion and, including deals still in negotiation, as high as about $750 billion. Positive quarterly cash flow does not retire a take-or-pay cloud book.

For someone underwriting either name, the cleaner current P&L is Anthropic’s. The larger consumer distribution and the more explicit multi-year funding hole are OpenAI’s. Neither profile is “strong” in a conventional software sense. One is profitable on an adjusted basis with a huge off-balance-sheet-style commitment stack; the other is still deeply cash-negative with a published plan that requires another nine-figure billions of capital before the decade is out.

Governance will be priced, not just disclosed

Both are public benefit corporations, but the control mechanics are different, and public investors will notice.

OpenAI’s October 2025 recapitalization put the business in OpenAI Group PBC and left the OpenAI Foundation in control: the Foundation holds about 26% of the equity and appoints the PBC board. Microsoft’s economic stake was about 27% as-converted at closing and later about 25% in its own annual report. Investors, including the new 2026 anchors, do not run the board. An April 2026 rewrite of the Microsoft deal capped revenue-share payments, made the IP license non-exclusive through 2032, and let OpenAI sell on any cloud. That reduced one governance-and-economics overhang. It did not turn OpenAI into a normal controlled company. Public shareholders would be buying a PBC whose directors are chosen by a nonprofit.

Anthropic’s structure is more unusual still, and it is about to be tested in a prospectus. The Long-Term Benefit Trust holds no meaningful economic stake and can elect a majority of the board—reported as four of seven—through Class T shares. Trustees are paid for time, not equity. The current group includes former Fed chair Ben Bernanke; on October 21 the company added Harvard physicist Christopher Stubbs. The Wall Street Journal has noted that the trust agreement itself has not been published, that the trust already shapes the board, and that no tech company has listed with this kind of non-shareholder override. Separately, Anthropic said on October 18 it would commit 5% of shares to a new Public Benefit Fund, with the board able to add 2% for every 50% increase in value above a baseline, and that combined with existing employee and founder pledges roughly 20% of shares are aimed at public-benefit uses. Reporting has also described a founder voting vehicle that would direct a Class F share so a majority of founders control 50.1% of votes on key matters, with public holders electing none of the board on their own. NYC Comptroller Mark Levine has already asked the company to pause the IPO over safety and governance. That letter will not stop a listing. It is a preview of the questions the roadshow will get.

Who lists first, and what the first print does to the other

Anthropic is the likely first listing, and by a wide margin if markets cooperate. It filed confidentially on June 1, a week before OpenAI’s June 8 submission. Sam Altman told Fortune a 2026 IPO would be ill-timed because of safety work and put the debut in 2027; CFO Sarah Friar told employees the company would be public in 2027, and that Anthropic going first was acceptable. Bankers split Goldman Sachs and Morgan Stanley into separate teams so the same houses can work both deals without sharing information—an arrangement the Wall Street Journal described as rare. Anthropic’s syndicate is further along operationally: a revolver expansion toward $15 billion, led by Morgan Stanley with Goldman, JPMorgan, and Citi in prominent roles, is the usual prelude to notifying banks of formal IPO roles.

The November window is narrow. Midterm elections, then Thanksgiving, then year-end. The Financial Times noted Polymarket pricing roughly a 60% chance of a November listing; CNBC cited Kalshi at about 75% that Anthropic is the last major 2026 IPO still likely to be officially announced this year. A slip into 2027 is the main alternative, not an OpenAI surprise debut.

On the financial profile that public investors can underwrite today, Anthropic is stronger. It has crossed into adjusted profitability and positive cash flow while still growing recognized revenue faster than almost any scaled software company on record. OpenAI has the better-known product, a larger private capital stack already closed ($122 billion in March), and a year-end run-rate target that, if hit, narrows the revenue gap. It does not have a credible near-term path to self-funding the plan it has shown investors. PitchBook’s October note put a fair-value sketch near $1 trillion against the corrected $50 billion run rate, below the $1.4 trillion talk—useful as a dissent, not as a market price.

The feedback loop runs both ways, but Anthropic moves first.

• A clean book at $1.8–2 trillion, with the stock holding after the first lockup-relevant weeks, validates frontier-lab multiples in the mid-20s to low-30s on run-rate sales and makes OpenAI’s private $1.4 trillion pre-money easier to defend into a 2027 listing. Amazon and Alphabet, whose Anthropic stakes are large enough to matter in their own earnings (Amazon carried $190.4 billion at June 30), get a mark-up. Microsoft, Nvidia, and SoftBank get a comp for OpenAI.
• A cut-price or volatile deal—the $1.5 trillion number some investor-day attendees were already discussing—resets the comp down before OpenAI ever files publicly. Reuters reported that OpenAI wants its round done before those disclosures invite exactly that comparison. Other issuers are already waiting out Anthropic so they are not competing for the same orders. A weak print would also pressure the circular-financing story: the cloud providers are simultaneously Anthropic’s largest capacity vendors, major shareholders, and, in Amazon’s case, an OpenAI investor too.
• Even a successful Anthropic listing does not clear OpenAI’s path. Public holders will have a traded pure-play with published margins, customer concentration, and a non-cancelable compute schedule. OpenAI will be marked against that every quarter it stays private. If Anthropic’s stock trades at a governance discount because of the Long-Term Benefit Trust, OpenAI’s Foundation-controlled PBC will be asked why its version should clear a higher multiple.

Public-market appetite: large, but no longer indiscriminate

Appetite is real and narrower than it was in the spring. SpaceX’s June listing, priced around $1.8 trillion, is the ceiling everyone is quoting. Anthropic is being discussed as the deal that could exceed it, with a primary raise talked about as high as about $100 billion. That only works if long-only funds and sovereigns treat a frontier lab as a scarce asset rather than as another capex story.

The pushback is already in the price talk, not just in commentary. After Q3 growth slowed from more than doubling quarter-on-quarter to about 45%, some existing and prospective investors argued for closer to $1.5 trillion than $2 trillion. Chip stocks sold off around the same investor meetings on inflation, higher yields, and worry about debt-financed data-center buildout. The New York Times framed Anthropic as the canary: a delayed or discounted debut would mark down stakes sitting on Amazon and Google’s books and, by comp, OpenAI. Safety language in the draft—catastrophic-risk warnings were reported from earlier prospectus reporting—plus an opaque trust that can remake the board, is a feature for some mission-aligned capital and a structural discount for others.

What would change the ranking: OpenAI hitting the $70 billion year-end run rate on its own accounting, with enterprise growth still above 100% and a visible narrowing of quarterly cash burn, would strengthen the 2027 case and could support the $1.4 trillion private mark even if Anthropic prices cautiously. Anthropic showing, in the public S-1, that Q4 monthly growth reaccelerated and that the non-cancelable cloud book is matched by contracted customer demand would support the high end of its range. Neither of those is in hand on October 22. What is in hand is a filing calendar that puts Anthropic in front of public investors within weeks, and a rival that has chosen to meet that moment with another private round rather than a prospectus.

---

Recent Findings Supplement (October 2026)

Anthropic is on track to test public-market appetite for frontier AI labs first with a November 2026 IPO (potentially at $1.5–2T), while OpenAI accelerates a large private round and targets 2027. Recent disclosures show both companies posting explosive revenue growth alongside massive infrastructure commitments and accounting-driven losses, setting up Anthropic’s debut as a direct gauge of whether investors will pay 20–40x annualized run rates for high-burn AI developers.[1][2]

Valuations and Recent Funding

OpenAI closed a record $122 billion round in March 2026 at an $852 billion post-money valuation (led by Amazon, NVIDIA, and SoftBank). It is now in advanced talks for $30 billion or more at a $1.4 trillion pre-money valuation, with UAE funds such as Abu Dhabi’s MGX in discussions to anchor the round. The company is racing to close this financing before Anthropic’s expected listing to lock in capital and strengthen its balance sheet ahead of a 2027 IPO.[3][4]

Anthropic raised $65 billion in May 2026 at a $965 billion post-money valuation. It confidentially filed for an IPO in June and is targeting a November 2026 debut (roadshow potentially the week of November 9, trading before Thanksgiving) at a valuation that investors have discussed in the $1.5–2 trillion range after recent disclosures, down from some earlier $2 trillion expectations.[1][5]

Side-by-side snapshot (latest reported figures):

• OpenAI: $852B post-money (Mar 2026); targeting $1.4T pre-money private round (Oct 2026 talks).
  
• Anthropic: $965B post-money (May 2026); IPO target discussed at $1.5–2T (Oct 2026 investor feedback).[6]

Implication: The compressed timeline for OpenAI’s private round reflects direct competition for overlapping investor capital; a strong Anthropic IPO could ease or tighten terms for OpenAI’s 2027 plans depending on multiples achieved.

Revenue Run Rates, Growth, and Burn

OpenAI’s annualized revenue run rate reached approximately $50 billion by the end of September 2026 (up from roughly $40 billion in August and ~$20 billion at end-2025), with Q3 growth of 77% overall and 107% in enterprise (now the majority of revenue). The company expects to reach or exceed $70 billion annualized by year-end 2026, driven by enterprise expansion. It reported ~$13 billion in 2025 revenue against a $38.5 billion net loss.[7][8]

Anthropic reported Q3 2026 revenue of $16.8 billion (up ~45–46% from Q2’s $11.5–11.6 billion and from Q1’s ~$4.7–4.8 billion). H1 2026 revenue totaled $16.3 billion (vs. $1.1 billion prior year). Its annualized run rate stood at ~$65 billion in late July (with some estimates at $74 billion later), with expectations of remaining below $100 billion by year-end. It described itself as adjusted profitable in Q3.[9][2]

Both companies carry heavy forward commitments: OpenAI faces projected cumulative burn approaching $280 billion by end-2030; Anthropic disclosed at least $518 billion in infrastructure commitments over roughly a decade (potentially up to $750 billion), with H1 2026 infrastructure spend nearing $9 billion.[6][10]

Implication: OpenAI shows slightly faster recent sequential acceleration in some metrics and a clearer path to the $70 billion+ run rate, but both exhibit similar patterns of revenue growth lagging behind multi-hundred-billion-dollar compute obligations. Public investors will scrutinize whether growth sustains at current multiples.

Investor Bases

OpenAI’s major backers include Microsoft (~25–27% stake post-recapitalization and recent rounds), Amazon (up to $50 billion commitment in the March round), NVIDIA, SoftBank, and discussions with UAE sovereign funds such as MGX. Overlap exists with Anthropic investors (e.g., Sequoia, Founders Fund).[11][12]

Anthropic’s largest disclosed investors include Amazon (largest single investor, ~$8 billion reported in some coverage) and Google (~$3 billion total). It has raised roughly $125–130 billion cumulatively, with significant overlap (~90 shared investors) alongside OpenAI backers.[13][14]

Implication: Shared investors create hedging dynamics; success or valuation compression in one IPO could influence allocations and terms for the other.

Governance Structures

OpenAI operates as OpenAI Group PBC (public benefit corporation) controlled by the OpenAI Foundation (nonprofit), which holds ~26% equity and retains sole power to appoint/remove the PBC board and enforce mission/safety decisions. Microsoft holds a minority economic stake without control.[15]

Anthropic is a PBC with a novel “Founder LLC” structure giving its seven co-founders 50.1% voting control via Class F stock on key matters. It recently committed 5% of shares (potentially scaling with valuation) to a Public Benefit Fund for AI safety, biomedical research, and related causes (plus founder/employee pledges bringing ~20% earmarked for public good overall).[16][17]

Implication: Both emphasize mission alignment via PBC status, but Anthropic’s founder voting bloc and explicit equity commitments to a benefit fund represent distinct mechanisms that may appeal differently to public investors focused on governance and societal impact.

IPO Preparations, Timing, and Market Appetite

OpenAI confidentially filed its S-1 in June 2026; lead banks referenced in coverage include Goldman Sachs and Morgan Stanley. CEO Sam Altman has ruled out a 2026 listing, with targets now pointing to 2027. The company is accelerating its private round partly to prepare its balance sheet.[18]

Anthropic confidentially filed in June 2026, with Goldman Sachs and Morgan Stanley as lead underwriters (JPMorgan and Citi also involved). It plans a public S-1 filing as early as October 23, 2026, targeting a November listing—positioning it as the first major test of AI-lab valuations post-SpaceX’s record IPO.[19][20]

Assessment: Anthropic is positioned to list first (November 2026 vs. OpenAI’s 2027 target). OpenAI currently shows a stronger near-term revenue trajectory and enterprise momentum, but both carry comparable high-burn profiles. A successful Anthropic IPO at attractive multiples would likely boost sentiment and ease OpenAI’s path; a muted reception or valuation reset could pressure OpenAI’s private round terms and delay or lower expectations for its 2027 debut. Public appetite remains high for AI growth stories but is tempered by scrutiny of compute costs, accounting losses, customer concentration, and sustainability at 20–40x run-rate multiples.[2]

These developments are drawn exclusively from sources published after April 22, 2026.