NextFin News - Anthropic, the maker of the Claude chatbot, is on track to generate annualized revenue of more than $65 billion, according to people familiar with the matter, a run rate that has risen more than sevenfold from the end of last year and is reshaping the math behind what could become the largest initial public offering in history. The figure, which reflects performance through the end of July, puts the private AI lab ahead of rival OpenAI's reported $40 billion annualized pace and sets up a public-market test of whether enterprise demand for AI can sustain a valuation that bankers are pricing off revenue expected two years from now.
The Numbers Behind the Run Rate
The $65 billion annualized revenue is a run-rate figure — a projection of a full year's revenue extrapolated from a recent, shorter period — not GAAP trailing revenue. That distinction matters because the company's underlying quarterly disclosures show growth that is still accelerating. In the second quarter of 2026, Anthropic reported preliminary revenue of more than $11.5 billion, according to documents reviewed by people with knowledge of the figures. That is more than 14 times the $787 million recorded in the same quarter of 2025 and more than double the $4.73 billion posted in the first quarter of 2026.
The trajectory is steep even by the standards of the current AI investment cycle. At the end of 2025, the company's annualized pace was roughly $9 billion. By late May, after a $65 billion Series H funding round that valued the company at $965 billion post-money, Anthropic said its run rate had crossed $47 billion. Two months later it stood above $65 billion. At that pace, investors close to the company expect full-year 2026 revenue to land between $100 billion and $120 billion, and the company has told bankers and prospective investors that it is projecting revenue of roughly $190 billion to $200 billion by 2028.
That 2028 figure is the number doing the heavy lifting in the IPO valuation. At the $965 billion private mark, the current run rate implies roughly 15 times revenue. A public debut above $1 trillion would keep the multiple near that level against today's pace — but it would fall to about five times if the company delivers on the 2028 target. Bankers have been asked to price the listing using enterprise value-to-revenue multiples applied to forward forecasts, a practice with precedent in this cycle: Cerebras backers cited 2028 expectations ahead of its May debut, and SpaceX ran its projections to 2029 before its June listing.
The funding that preceded the filing underscores how much capital is required to keep this growth curve intact. The Series H round, closed May 28, 2026, brought together a syndicate led by Altimeter Capital, Dragoneer, Greenoaks and Sequoia Capital, with Coatue and ICONIQ among the co-leads, and included strategic infrastructure investors such as Micron, Samsung and SK Hynix. Roughly $15 billion of previously committed hyperscaler investment was rolled into the round, including $5 billion from Amazon. The capital is earmarked largely for compute expansion — the single largest line item on any AI lab's balance sheet — and for the long-dated infrastructure commitments that lock in chip supply but also convert a variable cost into a fixed obligation.
Why the Growth Is Accelerating Now
The question behind the $65 billion is not whether AI demand exists — hyperscaler capital spending has made that clear — but why Anthropic specifically is pulling ahead. The answer runs through three channels: enterprise distribution, developer tooling, and secured compute.
First, the revenue mix has shifted from consumer subscriptions toward enterprise deployments. Businesses are embedding Claude into coding workflows, research operations, and customer-facing automation, where contracts are larger, stickier, and priced on usage that compounds as adoption spreads inside an organization. The sequential doubling from $4.73 billion in the first quarter to more than $11.5 billion in the second suggests that enterprise contracts signed earlier in the year are now converting into billable usage at scale. Enterprise buyers also tend to expand usage within a quarter rather than renegotiate price, which makes the revenue stream more predictable than a consumer subscription book.
Second, coding and agentic workflows have become the wedge. Claude Code and related developer products have turned the model into infrastructure rather than a destination application. When a development team builds its workflow around a model, switching costs rise quickly — the model learns the codebase, the tooling conventions, and the deployment pipeline. That creates a retention profile closer to a cloud platform than to a consumer app, and it explains why the company can sustain usage-based growth even as consumer chatbot attention rotates between competitors.
Third, compute supply — the binding constraint for every AI lab — has been locked down through long-dated agreements. Anthropic has secured multi-gigawatt commitments from Amazon Web Services, Google and Broadcom, and other infrastructure partners. Those deals remove the risk that growth stalls because chips are unavailable, but they also convert a variable cost into a fixed, long-dated obligation. The company has projected its first quarterly operating profit, of roughly $559 million in the second quarter, on an adjusted basis — an early signal that usage growth is beginning to outpace the inference and cloud costs that have kept most AI labs in the red.
Anthropic said in its announcement that the confidential draft S-1 registration statement it submitted to the Securities and Exchange Commission "gives us the option to go public after the SEC completes its review," adding that any offering "will depend on market conditions and other factors."
The IPO Window: Record Deals, But a Reckoning on Multiples
Anthropic is preparing to enter an IPO market that has never been more receptive to AI — and never more demanding on proof. The first half of 2026 produced record venture-backed listing activity, led by SpaceX's $1.77 trillion debut in June, the largest initial public offering on record, and Cerebras' May listing, which priced at $185 and closed its first day up 68 percent after raising $5.55 billion as the largest semiconductor IPO ever.
Those debuts prove the window is open. They also set the comparables that will constrain Anthropic's pricing. Public AI and cloud names are trading at elevated revenue multiples — Palantir at roughly 53 times expected 2026 revenue, Cloudflare and SpaceX at about 41.6 times, according to market data compiled from public filings and index providers. At $965 billion of private value against a $65 billion run rate, Anthropic already sits near 15 times current revenue.
That is the central tension of the deal. Pricing off 2028 revenue is not unusual for hypergrowth software without a mature profit profile, but it moves the burden of proof forward two years. The market is being asked to underwrite a revenue path that more than triples from today's run rate, while the company simultaneously services multi-gigawatt compute commitments and navigates regulatory scrutiny over model use and exports.
The aftermarket record of 2026's listings adds a second layer of caution. Several of the year's high-profile debuts have struggled after their first-day pops, with shares trading well below opening levels as lockup expirations approached. That pattern matters for Anthropic because the investor base for a $1 trillion-plus listing is narrower than for a typical IPO — it requires large institutional mandates with the capacity and the risk tolerance to hold a position of that size. If the aftermarket sours before Anthropic prices, the company could face a choice between accepting a lower valuation or waiting for the window to reopen.
The Counter-Thesis: Run-Rate Math Can Break
The strongest case against the bull narrative is not that growth will stop — it is that run-rate figures systematically overstate durable revenue, and that the multiple compresses before the 2028 target is reached. A run rate extrapolates a hot quarter across twelve months; it assumes no churn, no price compression, and no slowdown in new bookings. History is littered with hypergrowth software companies whose run rates looked unstoppable six months before a reset.
There is also the margin question. Positive adjusted operating income is not the same as GAAP profitability, and the adjustment excludes stock-based compensation and other items that public-market investors watch closely. If inference costs rise faster than pricing power — as competitors drive down API prices — the path from $65 billion of revenue to $559 million of quarterly operating profit could narrow rather than widen. Several analysts have flagged that gross margin after cloud and inference costs remains the decisive unknown in AI-lab valuation, and that number will not be fully visible until the formal S-1 becomes public.
Then there is the competitive overlay. OpenAI's annualized revenue has reportedly reached about $40 billion, and the two companies may calculate their metrics differently. If the gap between them closes, or if a competitor lands a comparable enterprise distribution deal, the growth premium embedded in Anthropic's private valuation has less room to hold. The counter-thesis, in one line: the market is paying for a 2028 destination while the company is still proving the durability of its 2026 velocity.
Regulatory Overhang: The Pentagon Designation
One risk sits outside the revenue model entirely. In March 2026, the U.S. Department of Defense designated Anthropic a supply-chain risk — the first such designation ever applied to an American company — after the administration directed federal agencies to phase out the company's technology. Anthropic filed lawsuits in two federal courts challenging the designation, and a federal judge later indicated she was likely to block the government from applying the label. The dispute remains unresolved, and the company has warned that it could jeopardize billions of dollars in potential government revenue.
The designation matters for the IPO because it introduces a binary policy risk that cannot be hedged with a revenue forecast. Even if Anthropic prevails in court, the episode signals to government customers — and to the enterprise clients that watch Washington closely — that the company's relationship with the state is contested. That can slow procurement cycles and give rivals an opening in the public-sector segment, where contract awards are large but politically sensitive.
What Would Prove the Bull Case Wrong
The falsifying signal is specific and observable: if Anthropic's quarterly revenue growth decelerates to below 50 percent year-over-year for two consecutive quarters after the IPO filing, or if gross margin disclosed in the S-1 comes in below 60 percent, the structural-growth thesis is broken. Either outcome would indicate that the run rate was capturing a transient surge rather than a durable enterprise platform, and the 2028 bridge would need to be re-priced.
What to Watch and What It Means
Short term (sentiment and liquidity): The IPO timing remains officially undecided. The company has not set a share count, price, or listing date, and a spokesperson said in December 2025 that it had "not decided when or even if" it would go public. Market participants have pointed to a potential autumn 2026 debut, but that timing is contingent on equity-market conditions and the SEC review of the confidential filing. Any volatility in the AI trade — a pullback in Nvidia, Broadcom, or the hyperscalers — would narrow the window quickly.
Medium term (fundamentals): The formal S-1 will be the first audited look at revenue quality, customer concentration, gross margin, and the size of the long-dated compute obligations. Investors should watch three lines: revenue recognized versus billings, the top-customer concentration, and the cash cost of inference per query. The second-half 2026 bookings trend will determine whether the $100 billion-to-$120 billion full-year expectation holds.
Long term (structural): The base case is that Anthropic lists in late 2026 at a valuation between $1 trillion and $1.5 trillion, with the 2028 revenue target serving as the anchor for institutional demand. The upside case is that enterprise AI adoption compounds faster than expected, gross margins expand as inference costs fall, and the company re-rates toward the 40-times-forward-revenue range of its closest comparables — implying a market capitalization well above $2 trillion within two years of listing. The downside case is that growth decelerates, margins compress, and the stock trades down to the 15-to-20-times range typical of maturing software, which at a $1 trillion entry point would mean a significant drawdown for public investors.
Who benefits and who is exposed: a successful, well-received Anthropic listing would validate the entire private AI valuation complex, lifting late-stage AI names and the hyperscalers that host them. The exposed are public-market investors buying at the top of the multiple range, and any AI company whose valuation rests on a 2028 revenue promise without the quarterly execution to match it.
The $65 billion run rate is real, but it is a speedometer, not a map. Anthropic's IPO will test whether the market is willing to pay today for a 2028 destination — and whether a company can grow fast enough to outrun the multiple it was born with.
Explore more exclusive insights at nextfin.ai.
