Everything About AI
NextFin News - Anthropic is close to handing the top roles in what could become the largest initial public offering in history to Morgan Stanley and Goldman Sachs, with a listing that investors and bankers are discussing at a valuation near $2 trillion and a launch window as soon as October 2026. The move would more than double the $965 billion price tag the AI lab commanded when it closed a $65 billion Series H round on May 28, and it sets up a direct collision with rival OpenAI, which is racing toward its own debut. The central question is not whether Anthropic can list — a confidential draft registration statement was filed with the Securities and Exchange Commission on June 1 — but whether public markets will pay roughly 31 times an annualized revenue run rate that, however explosive, has existed for less than a year.
The Deal Takes Shape
The selection of lead underwriters is the clearest signal yet that Anthropic intends to move from private fundraising to a public listing within months rather than quarters. Morgan Stanley and Goldman Sachs are being lined up for the top roles, with JPMorgan Chase also working on the offering. The company has already submitted a confidential draft S-1, a procedure that lets issuers negotiate the prospectus with regulators away from public view before unveiling the filing — the same path SpaceX, OpenAI and a growing roster of 2026 technology candidates have taken.
The numbers framing the transaction are striking. At the $2 trillion valuation under discussion, Anthropic would price at roughly 31 times its $65 billion annualized revenue run rate as of the end of July 2026. That run rate is the product of an extraordinary acceleration: the company was generating about $9 billion annualized at the end of 2025, climbed to $47 billion by May, and passed $65 billion by the end of July — a sevenfold increase inside a single year. Second-quarter revenue alone exceeded $11.5 billion, and adjusted operating income turned positive for the first time, according to figures the company has shared with investors.
The $2 trillion figure is not pulled from thin air: it rests on a projected 2028 revenue range of $190 billion to $200 billion, which would put the listing at roughly 10 times forward sales — a far more defensible multiple than the 31 times trailing revenue. But that framing requires the company to roughly triple its revenue in the next two years, and to keep doing it after the easy enterprise land-grab is over.
Even so, the jump from a $965 billion private mark to a $2 trillion public one is not a rounding error. It implies that public-market investors would be asked to assign a premium of more than 100 percent over the price that Altimeter Capital, Dragoneer, Greenoaks and Sequoia Capital paid in May — and to do it for a company whose path to sustained profitability is only just coming into view.
The Multiple That Has to Be Defended
Every IPO price is a story about the future, but the $2 trillion Anthropic story has to clear a very specific bar: approximately 31 times sales on a trailing basis. For comparison, OpenAI — the natural peer — is valued at $852 billion against a reported run rate above $40 billion, which prices it at roughly 21 times sales. A $1 trillion OpenAI listing, as some bankers have discussed, would take that to 25 times. Anthropic at $2 trillion would sit meaningfully above both.
The public-market yardstick is harsher still. Microsoft trades at about 11 times sales; Nvidia, the engine of the entire AI buildout, at roughly 22 times. No company has ever gone public at a $2 trillion valuation, and the only technology issuer in Anthropic's weight class in recent memory, SpaceX, listed in June 2026 at an IPO valuation near $1 trillion before the market rerated it higher — to about $2.59 trillion — after months of public trading, audited financials and a visible order book.
That sequence matters. SpaceX did not debut at $2.59 trillion. It earned the rerating in the open market, where buyers could inspect actual filings rather than investor-update slides. Anthropic would be attempting the reverse: arriving at a frontier valuation on day one, before the public has seen a single audited quarter. The burden of proof sits with the issuer, not the buyer.
Is the $65 Billion Run Rate Durable?
The bullish case rests on the revenue trajectory, and the trajectory is real. Anthropic's growth has been powered by an enterprise-first go-to-market built around the Claude model family and Claude Code, the AI coding assistant that has become the fastest-growing product in the company's lineup. Roughly 80 percent of Anthropic's revenue comes from business customers, and enterprise contracts carry multi-year commitments, higher retention and expanding seat counts — the kind of revenue quality that supports premium multiples.
There is, however, a mechanical caveat that every buyer will scrutinize. A substantial share of Anthropic's top line flows through cloud partnerships, where the company resells model access via Amazon Web Services, Google Cloud and Microsoft Azure. Under this arrangement, gross customer spend is booked as revenue with the cloud partner's cut recorded as a cost — which inflates the reported top line relative to a company that reports net of partner fees. Anthropic sells a meaningful share of Claude access through AWS Bedrock and Google Cloud's Vertex AI and reports those sales gross, while OpenAI's ChatGPT and API revenue flows partly through its commercial arrangement with Microsoft and is reported net — a structural accounting difference that can swing headline ARR by 20 to 30 percent for otherwise comparable businesses. The distinction does not change the direction of travel, but it means the $65 billion run rate and a clean subscription number are not the same animal. Once the S-1 opens the books, the market will separate the two, and the multiple will be set on the quality of revenue that remains.
The profitability picture is improving but still early. Adjusted operating income turned positive in the second quarter, yet the company remains in heavy investment mode: it has signed agreements with Amazon for up to five gigawatts of new compute capacity, with Google and Broadcom for five gigawatts of next-generation TPU capacity, and with SpaceX for GPU access to the Colossus 1 and Colossus 2 clusters. Compute commitments of this scale are a growth asset only if demand keeps compounding. If enterprise spending on AI slows, the same capacity becomes a fixed-cost anchor.
The Cap Table: Who Wins, Who Is Exposed
An IPO at $2 trillion would create paper gains on a scale rarely seen in venture history, and the distribution is concentrated. Google holds a 14 percent stake — the only externally verified ownership figure, disclosed in a 2025 antitrust court filing — worth roughly $280 billion at the discussed valuation. Amazon is the largest single investor, with more than $8 billion committed across rounds from 2023 through 2026 and an estimated mid-teens percentage. Microsoft and Nvidia joined the cap table in November 2025 with commitments of up to $5 billion and up to $10 billion respectively, and Salesforce holds a position that was worth about $5 billion in June 2026.
For the hyperscalers, the investment is strategic as much as financial: Anthropic is the first frontier model available across all three major clouds — AWS, Google Cloud and Azure — and each partner gains a reason for enterprise customers to build on its platform. That alignment is also the source of the listing's biggest structural tension. Anthropic is a public benefit corporation, legally bound to a safety mission that can conflict with partner demands for model capability and speed. The founders, Dario and Daniela Amodei and their co-founders, retain meaningful equity and governance control designed to keep the safety mandate intact. Public shareholders will be buying into a company whose fiduciary duties are split — a structure that commands a discount in some corners of the market and a premium in others, depending on whether investors view the safety mission as a moat or a constraint.
The IPO Wave and the Index Problem
Anthropic is not listing into a vacuum. SpaceX went public on June 12, 2026, pricing at $135 and closing its first day at $162 before climbing to roughly $201.80, a market capitalization of about $2.59 trillion. OpenAI filed its own confidential S-1 on June 8, and its chief financial officer, Sarah Friar, has signaled that Anthropic could move first.
"As you know we are confidentially under file, and Anthropic is also under file. There is a chance they pull the cover off that confidential file in the coming weeks and become public in September. That's OK, we are running our own race." — Sarah Friar, chief financial officer of OpenAI
The sequencing matters because of a rule most retail investors do not see. Major indexes require a seasoning period — roughly twelve months of public trading history — before a newly listed mega-cap qualifies for inclusion in the S&P 500. For at least a year after listing, index funds will not hold a single share of Anthropic, no matter how large it becomes. Demand on debut will therefore come from active managers, sovereign wealth funds and momentum capital, not from passive index flows. That makes the first-day price more volatile and more dependent on the quality of the roadshow narrative.
There is also a crowding question. Goldman Sachs expects total equity issuance in 2026 to reach about $600 billion including IPOs — still less than 1 percent of US equity market capitalization, which suggests the broader market can absorb these deals without major dislocation. But within the technology sector, the effect is more concentrated: analysts project the sector's weight in the S&P 500 could climb from roughly 51 percent to about 54 percent once SpaceX, OpenAI and Anthropic are fully included. For context, technology peaked at about 35 percent of the index during the dot-com bubble.
The Counter-Thesis
The strongest argument against the $2 trillion number is simple: private-market pricing has overshot, and the IPO will be the moment of correction. Secondary-market platforms have already pushed Anthropic's implied value to about $1.2 trillion — a 550 percent year-over-year increase — yet actual trades remain rare because holders are not selling. A quoted secondary price in a thin market is not the same as a cleared auction, and the spread between the $965 billion primary mark, the $1.2 trillion secondary quote and the $2 trillion IPO discussion is itself a warning sign that price discovery has not happened.
This counter-thesis has force, but it rests on a premise that the revenue data undermines. A deep correction from $2 trillion would require the market to conclude that Anthropic's $65 billion run rate is largely illusory — that enterprise demand is pulling forward, that churn is hidden inside gross reporting, or that the AI spending cycle is about to stall. None of those conditions is visible today. Enterprise AI adoption is a structural shift, not a cyclical spike: model usage is embedded in software development workflows, customer-service stacks and knowledge-management systems that are expensive to rip out. The multiple may be cyclical; the revenue base underneath it is not.
The falsifying signal is specific: if Anthropic's reported revenue growth decelerates to below 30 percent year over year for two consecutive quarters after the S-1 is unveiled, or if gross margins on the cloud-partner revenue stream come in materially below the 50 percent level that a pure software business would command, the structural-growth thesis breaks and the $2 trillion figure becomes indefensible. Until then, the burden sits with the skeptics to prove that the demand is fake, not with the company to prove that it is real.
What Comes Next
Short term (the listing window): The base case is a fourth-quarter debut, most likely October, with the confidential filing going public and a roadshow priced off the July $65 billion run rate. The upside case is a September pull-forward if OpenAI's timeline forces Anthropic's hand; the downside case is a slip into early 2027 if equity markets weaken or if the S-1 reveals margin or concentration risks that reset banker models.
Medium term (the first four quarters): The key metric is not the debut pop but the post-listing rerating path. SpaceX showed that a mega-IPO can climb after listing if the filings validate the story; the reverse is equally possible. Watch quarterly revenue growth, the gross-versus-net revenue split, and whether adjusted operating income stays positive as compute depreciation ramps.
Long term (structural): If Anthropic sustains enterprise-grade growth and keeps its safety mandate from becoming a governance drag, it can grow into the multiple over several years — the way Nvidia grew into its own frontier valuation through earnings rather than multiple expansion. If the AI spending cycle proves cyclical after all, the $2 trillion tag will be remembered as the peak of the private-market bubble, and the public shareholders who buy at debut will underwrite the hangover.
The deal is not done, the price is not set, and the S-1 has not been unveiled. But the selection of Morgan Stanley and Goldman Sachs tells the market what to expect: the largest IPO in history is no longer a question of if, but of how much investors are willing to pay for the future of artificial intelligence on day one.
Explore more exclusive insights at nextfin.ai.
