Everything About AI
NextFin News - In February 2026, Anthropic's chief executive warned that missing AI demand forecasts by a single year could push the company into bankruptcy. Three months later, the Claude maker had grown 80-fold in a single quarter, closed a $65 billion funding round at a $965 billion valuation, and quietly filed a confidential IPO registration with the SEC. The question no longer is whether Anthropic survives; it is whether a company that went from roughly $100 million in annualized revenue at the start of 2025 to more than $65 billion by July can justify a public valuation that could exceed $1.75 trillion.
The IPO That Redraws the Tech Pecking Order
Anthropic confirmed on June 1, 2026 that it had confidentially submitted a draft registration statement on Form S-1 to the U.S. Securities and Exchange Commission for a proposed initial public offering of its common stock. The filing, reported by multiple outlets, puts the company behind Claude on a path to go public as early as the fall of 2026, with some reports targeting a valuation between $1.75 trillion and $2 trillion and a raise of up to $75 billion — potentially the largest IPO in history if those targets hold.
The revenue trajectory is what makes the filing credible rather than aspirational. Anthropic's annualized revenue run rate — a metric that projects full-year sales from a shorter period — climbed from roughly $100 million at the start of 2025 to $4 billion by July 2025, more than $9 billion at the end of 2025, about $20 billion by March 2026, $30 billion by April, more than $47 billion by early May, and past $65 billion by the end of July. That is a climb from $9 billion to $65 billion in seven months, a pace with no clean precedent among public-market software companies.
The valuation math is steeper still. In February 2026, Anthropic raised $30 billion at a $380 billion valuation. In May, it closed a $65 billion Series H round at $965 billion — surpassing rival OpenAI and making it the most valuable private company in the world, ahead of SpaceX's reported $1.77 trillion mark. Now the company is reportedly seeking a public valuation of $1.75 trillion to $2 trillion, more than double its last private round. For context, that would place Anthropic above nearly every listed technology company and at a price-to-sales multiple that leaves almost no room for execution error.
But the headline valuation is not the real story. The real story is what happened in the second half of 2025 that turned Anthropic from a well-funded also-ran into the presumptive front-runner in the AI race — and whether the market is paying for a durable structural advantage or simply extrapolating a wave that may be partly cyclical.
What Changed: The Claude Code Inflection Point
For most of the AI boom, OpenAI appeared to have already won. ChatGPT was the consumer brand, GPT-4 the developer default, and Microsoft's Azure the distribution engine. Anthropic, founded by former OpenAI researchers including CEO Dario Amodei, was respected for its safety-first posture but trailed on adoption and revenue.
The turning point came in November 2025 with the launch of Claude Code, an agentic coding assistant that let developers describe an application, website, dashboard, or research problem in plain English and have the AI complete the task. The product crossed $1 billion in annualized revenue within six months — the fastest any enterprise software product has reached that milestone — and by February 2026 enterprise usage accounted for more than half of Anthropic's total revenue.
The release of Claude Opus 4.5 in late 2025 compounded the momentum. The model became a hit among developers and enthusiasts, and it gave Anthropic a credible performance lead over OpenAI's offerings at a moment when ChatGPT's growth began to plateau. By early 2026, Anthropic's revenue growth had surpassed OpenAI's, and so had its valuation.
At Anthropic's Code with Claude developer conference in San Francisco on May 6, 2026, Amodei disclosed that the company's revenue and usage had grown 80-fold on an annualized basis in the first quarter, versus the company's internal plan for 10x growth.
"It's just crazy," Amodei told the crowd, adding that the growth was "too hard to handle" — a candid admission that the company's own forecasting models, calibrated for a 10x annual curve, had been rendered obsolete in weeks. The 80-fold figure is the CEO's statement about the first quarter; the independently reported run-rate path — $9 billion at the end of 2025 to more than $65 billion by July — is what anchors the IPO math.
That inflection is the mechanism behind the IPO. It is not that Anthropic found a better chatbot; it is that it found the first genuinely enterprise-paying AI workload. Coding is measurable, billable, and defensible — a developer either ships faster with Claude Code or does not. That measurability converted AI hype into invoiceable revenue faster than any competitor managed.
The Math the Market Is Being Asked to Underwrite
At a $1.75 trillion valuation against a $65 billion annualized revenue run rate, Anthropic would trade at roughly 27 times run-rate sales. Even at the more conservative $965 billion private-round valuation, the multiple is about 15 times — a level that assumes growth continues at a pace that has never been sustained by a company of that size.
The margin picture is improving but remains thin. Compute cost per dollar of revenue fell from 71 cents in early 2026 to 56 cents by mid-2026, according to public remarks by Chief Operating Officer Daniela Amodei. That efficiency gain matters: every 10-cent improvement in compute cost per dollar of revenue translates into billions of dollars of incremental gross margin at this scale. But a company spending 56 cents of every revenue dollar on inference alone is still operating with a gross margin structure closer to a capital-intensive utility than to high-margin software.
And the capital commitments required to stay in the race are staggering. Anthropic has signed agreements with Amazon for up to five gigawatts of new cloud capacity, with Google and Broadcom for five gigawatts of next-generation TPU capacity, and with SpaceX for access to GPU capacity in its Colossus 1 and Colossus 2 data centers. Amazon is reportedly committed to invest up to $25 billion in Anthropic, including $5 billion up front and another $20 billion tied to commercial milestones, while Alphabet has committed up to $40 billion, including $10 billion in cash at a $350 billion valuation and another $30 billion tied to performance targets.
Those hyperscaler commitments are both a lifeline and a leash. They guarantee Anthropic the compute it needs to scale; they also bind the company to specific cloud ecosystems as its two largest investors become its two largest customers and infrastructure providers.
Who Wins Before a Single Share Trades
An IPO is usually an exit for early backers. Anthropic's is different: its two most important backers are not venture funds looking to cash out — they are the cloud platforms Anthropic depends on to exist.
Amazon and Alphabet hold substantial stakes in Anthropic, and both marked those investments to market at the end of the second quarter against the $965 billion funding round. If the IPO doubles the company's value from its last private raise, both companies are positioned to book tens of billions of additional unrealized gains on the stock and convertible notes they hold. That accounting gain does not put cash in their pockets on day one, but it strengthens balance sheets, supports credit ratings, and — most importantly — justifies the next round of data-center spending to shareholders who have grown impatient with AI capital expenditure that lacks visible returns.
The spillover extends further. Broadcom sits on the other end of the five-gigawatt TPU commitment with Google, turning the Anthropic deal into a multi-year revenue stream for the chip designer. SpaceX earns capacity fees from its Colossus data centers while competing with Anthropic for its own place in the IPO queue. Microsoft, which quietly became one of Anthropic's largest customers at roughly $500 million a year in model spend, gains a hedge against its OpenAI dependence. The Anthropic IPO, in other words, is a wealth event for the entire AI supply chain before a single public investor owns a share.
That distribution of winners matters because it creates a coalition with an interest in the offering's success. The underwriters are not the only ones working to ensure the IPO prices well; the companies whose own earnings now depend on Anthropic's marked value have every incentive to see the debut land firmly.
Cyclical Wave or Structural Shift? The Core Call
Here is the judgment the IPO forces investors to make: is Anthropic's rise a cyclical wave that will mean-revert, or a structural regime shift that will not self-correct?
The cyclical case is uncomfortable but real. First, the 80-fold quarter rode a one-time product inflection — the launch of a category-defining tool into a market with almost no supply. No enterprise software company has ever sustained 80x growth; the law of large numbers alone makes reversion near-certain. Second, the AI infrastructure buildout is a classic capital-expenditure supercycle: hyperscalers are over-ordering chips and power today against demand that may not materialize at the same pace, and history suggests these cycles overshoot before normalizing. Third, the competitive response is already underway — OpenAI has declared a "code red" internally, GPT-5 shipped after a two-year wait, and Microsoft quietly diversified its model spend while hedging across providers. When the laggards catch up, pricing power erodes.
The structural case, however, is stronger. Three things about Anthropic's position look durable rather than cyclical. First, the moat is in the workflow, not the model: Claude Code embeds into the software-development lifecycle, and switching costs in an engineering organization — retraining, rewriting prompts, revalidating outputs — are high once a team standardizes. Second, the revenue quality has shifted from consumer subscriptions to enterprise contracts, which are stickier and less prone to the viral-attention whiplash that hit consumer AI products. Third, the compute commitments create a capacity barrier: five gigawatts with Amazon and five with Google and Broadcom represent infrastructure that competitors cannot replicate quickly, locking Anthropic into a supply position that smaller rivals cannot match.
The verdict: this is a structural shift wrapped in a cyclical spike. The 80-fold quarter is the cyclical leg — it will revert, and anyone underwriting 80x growth into perpetuity will lose money. But the underlying regime change — AI moving from experimentation to billable enterprise workflow, with Anthropic as the early leader in the most monetizable vertical — is structural and will not self-correct. The market's mistake would be to price the spike as the baseline. The investor's mistake would be to dismiss the company because the spike cannot last.
The Second-Order Trade Everyone Is Missing
The first-order read of the Anthropic IPO is obvious: a huge listing, a windfall for early investors, and a new megacap for index funds to absorb. The second-order consequence is what matters more.
When Anthropic prices at $1.75 trillion, it does not just set its own valuation — it re-rates the entire AI complex. OpenAI, SpaceX, and every private AI lab suddenly has a public-market anchor that validates higher private marks. That lifts the marked-to-market value of the stakes held by the hyperscalers that backed them. Alphabet and Amazon, which hold substantial stakes in Anthropic and marked their investments to market at the end of the second quarter against the $965 billion round, could book tens of billions of additional unrealized gains if the IPO doubles the company's value from its last private raise.
That creates a feedback loop the market has not fully priced: the more valuable Anthropic becomes at IPO, the stronger the balance sheets of its backers look, and the more capacity those backers can justify building — which in turn keeps Anthropic supplied with the compute it needs to keep growing. The IPO is not an exit event; it is the refinancing of the entire AI infrastructure buildout through public markets.
There is also a cross-asset implication. A $75 billion IPO raise absorbs liquidity at a moment when rate-cut expectations are already being repriced. If the offering lands as the Federal Reserve is easing, the market can digest it. If it lands into a risk-off backdrop, the sheer size of the float could crowd out other listings and pressure high-multiple growth stocks broadly. The IPO's timing is not just Anthropic's problem — it is a liquidity test for the entire 2026 tech issuance market.
The Strongest Case Against the Bull Thesis
The strongest counter-thesis attacks the valuation at its foundation: Anthropic is being priced as a software company while its cost structure is that of a capital-intensive infrastructure business. At 56 cents of compute cost per revenue dollar, gross margins are structurally capped well below the 75%–80% that public software markets reward. If the market re-rates Anthropic on a utility-like multiple rather than a software multiple, the $1.75 trillion target collapses toward something closer to the $965 billion private round — a roughly 45% downside from the reported IPO target.
This view has mainstream backing among value-oriented investors and analysts who argue that hyperscaler concentration is a structural risk, not a strength: Amazon and Alphabet are simultaneously Anthropic's largest investors, largest cloud providers, and potential future competitors with their own models. If either hyperscaler decides to prioritize its in-house AI over Anthropic's, the company's supply and distribution advantages evaporate overnight. The bankruptcy warning Amodei himself issued in February 2026 is Exhibit A for the fragility of the model: a company that was one year away from insolvency is not a compounder with a wide moat — it is a leveraged bet on a single product cycle.
The answer to the counter-thesis is that the market is not being asked to underwrite 80x growth forever — only long enough for the enterprise base to lock in. If Claude Code's installed base survives the first competitive wave and compute costs keep falling toward 40 cents per dollar of revenue, the margin structure improves faster than the skeptics assume, and the software multiple holds. But that is a big if, and it names the falsifying signal.
The signal that would prove the structural thesis wrong: if Anthropic's annualized revenue growth falls below 20% year over year within four quarters of the IPO while compute cost per revenue dollar remains above 50 cents, the structural-shift call fails — that combination would show the 80x quarter was a one-time product spike and that the cost structure cannot support software-like margins.
What Comes Next: Scenarios and What to Watch
Short term (liquidity and sentiment): The IPO is expected to price by the end of 2026. If it lands in a risk-on window with the Fed easing, the stock could open well above the $1.75 trillion target as index funds and momentum capital chase the only pure-play AI megacap available. If it lands into a risk-off backdrop, expect a muted debut and pressure on other high-multiple tech names as the $75 billion float absorbs liquidity.
Medium term (fundamentals): The public S-1 will disclose customer concentration, the share of revenue from Claude Code, and the terms of the hyperscaler agreements. Base case: revenue growth normalizes to 40%–60% annually as the enterprise base expands, and the stock holds near the IPO price. Upside case: Claude Code becomes the default engineering tool across Fortune 500 developers, growth stays above 60%, and the stock re-rates toward $2 trillion. Downside case: customer concentration proves severe, OpenAI's competitive response gains traction, and the stock falls back toward the $965 billion private valuation.
Long term (structural): The winners are the entities positioned on both sides of the IPO. Alphabet and Amazon hold stakes that could generate tens of billions in unrealized gains; Broadcom and the infrastructure supply chain benefit from the five-gigawatt TPU commitment; and the broader AI ecosystem gets a public-market anchor that validates higher valuations across the board. The exposed are the late-arriving private investors buying at the top of the round, and any high-multiple tech name that gets crowded out by the IPO's liquidity demand.
What to watch: the public S-1's customer-concentration table, the disclosed gross margin after compute costs, the IPO pricing relative to the $1.75 trillion target, and the Federal Reserve's rate path into the offering window. Each of these will determine whether Anthropic's IPO is remembered as the moment AI grew up — or as the moment the market paid for a growth spike that had already peaked.
The central judgment: Anthropic's IPO is less a company listing than a referendum on whether AI has crossed from hype into cash-generating infrastructure. The 80-fold quarter will not last, but the workflow it created just might — and the market is about to learn the difference at a $1.75 trillion price tag.
Explore more exclusive insights at nextfin.ai.
