The Maker of Claude Hit a $65 Billion Revenue Pace. It's Still Not on the Market. Yet

Key Takeaways

• The fastest revenue story in business history isn't on the stock market. Anthropic, the AI company behind the Claude models, has gone from a reported $1 billion annualised revenue pace in early 2025 to a reported $65 billion pace by July 2026, and it remains private, for now.

• The "earnings" everyone quotes aren't earnings: as a private company Anthropic publishes no quarterly reports. The figures in circulation are reported run-rates and funding-round disclosures, Bloomberg and the Wall Street Journal put preliminary second-quarter revenue above $11.5 billion, ahead of rival OpenAI's reported $6.7 billion, with Anthropic showing an adjusted operating PROFIT while OpenAI reported a large loss. All reported, none of it audited public filings yet.

• The valuation ladder has been vertical: a $183 billion round in September 2025, $380 billion in February 2026, $965 billion in May's $65 billion Series H, and, per the Financial Times and Wall Street Journal, investors discussing an IPO that could target around $2 TRILLION. Reuters has reported a confidential IPO filing; nothing is scheduled or guaranteed.

• You can't buy Anthropic today, but its story touches stocks you CAN research: Amazon (a reported $8 billion investor) and Alphabet hold stakes, Microsoft and Nvidia have reported investment commitments, and the chips it runs on come from the TSMC-ASML supply chain this blog covered all week.

• The honest catch: private-market figures are reported, not independently audited disclosures; trackers disagree at the edges; the implied revenue multiples (about 15 times) are generous by any historical standard; and this very week's hot-inflation selloff showed how fast AI-linked valuations can be repriced. Tickers named are examples to research, not recommendations; this guide is educational, not investment advice.

• Research it your way: you can explore global stocks and ETFs from just $1 with zero commission on the Nemo.money app.

Introduction

People keep searching for "Anthropic earnings", and the search contains a misunderstanding worth clearing up first: there are no Anthropic earnings. The company behind the Claude AI models is private, it files no quarterly reports, holds no earnings calls, and publishes no audited income statement for the public.

And yet the numbers attached to it, all from reporting by Bloomberg, the Wall Street Journal, Reuters, CNBC and the Financial Times, describe possibly the fastest revenue ascent any company has ever recorded: a roughly $1 billion annualised pace in early 2025, $30 billion by April 2026, $47 billion disclosed with May's funding round, and about $65 billion by late July. One funding round in May valued the company at $965 billion, nearly triple its February valuation, and reports since suggest a confidentially filed IPO that investors have discussed at around $2 trillion.

If those reports hold, the biggest listing in stock-market history is somewhere between rumour and runway, in the very month GCash listed in Manila, Jio opened India's biggest-ever IPO, and Bumble exited the market entirely. The IPO door has never been busier, and this would be its loudest knock ever.

This guide does what the hype can't: separates what's reported from what's filed, teaches the run-rate trick that makes $4.6 billion and $65 billion both true at once, and maps how an everyday investor can research the story TODAY, through the listed giants funding it and the chipmakers building it, without touching a single private share.

It's educational, not investment advice. To explore, you can invest in global stocks and ETFs from just $1 with zero commission on the Nemo.money app.

People are searching:

• "Anthropic earnings"

• "Anthropic IPO"

• "Anthropic stock"

• "Anthropic valuation"

• "Claude AI company"

• "how to invest in Anthropic"

• "Anthropic vs OpenAI revenue"

The Numbers, As Reported

Every figure below is reporting, not an audited public filing, sources named, because that distinction is half this article's point:

• 🚀 The revenue ladder (annualised run-rates, as reported): about $1 billion in early 2025 (CNBC), $30 billion by April 2026, $47 billion disclosed alongside May's funding round (CNBC), and roughly $65 billion by late July (Bloomberg; the tracker Sacra estimates similarly, with others ranging up to the low $70s billions, trackers disagree, which tells you something about private-market precision).

• 🤖 The engine, reportedly, is code: Claude Code, the company's AI coding assistant, went from a reported $1 billion annualised pace in late 2025 to $2.5 billion by February 2026 and kept doubling (Sacra), the enterprise coding boom in one product line.

• ⚔️ The rivalry scoreboard (reported, preliminary): Bloomberg and the Wall Street Journal put Anthropic's second-quarter 2026 revenue above $11.5 billion against OpenAI's $6.7 billion, with Anthropic reporting an ADJUSTED operating profit of about $559 million while OpenAI reported a $12.3 billion loss. If those hold, the number-two AI lab by fame became number one by revenue, quietly.

• 📅 And the anchor that keeps everyone honest: per Reuters' reporting on the company's IPO prospectus, Anthropic's actual CALENDAR-2025 revenue was about $4.6 billion. Not $65 billion, $4.6 billion. Both numbers are real; the gap between them is the next section's whole lesson.

The Valuation Ladder

• 🪜 Four rungs in nine months, as reported: $183 billion (September 2025 round), $380 billion (February 2026's $30 billion Series G, led by GIC and Coatue), $965 billion (May 2026's $65 billion Series H, led by Altimeter, Dragoneer, Greenoaks and Sequoia, per CNBC the moment it overtook OpenAI as the most valuable AI startup), and now, per the Financial Times and Wall Street Journal, investor discussions of an IPO around $2 TRILLION, with Bloomberg reporting talk of a roughly $100 billion raise. The last rung is rumour; the first three are closed rounds.

• 🧮 What the price implies: at $965 billion against a reported $65 billion run-rate, the market is paying roughly 15 times revenue, which one analysis notes is actually LOWER than the multiple implied by OpenAI's reported valuation, the "cheaper" of two extraordinary numbers, with total reported funding raised in the region of $125 billion.

• 🏰 Who's already in: the reported cap table reads like the market itself, Amazon (about $8 billion invested), Alphabet, and 2026 reporting of commitments from Nvidia (up to $10 billion) and Microsoft (up to $5 billion) alongside a large Azure compute deal, plus sovereign and institutional names from GIC to Fidelity. Remember what that means mechanically: when a private company's value rises, its EXISTING listed shareholders quietly mark up too.

• ⚠️ And what a ladder like this is not: proof. Each rung is a negotiated price between a company and willing buyers at one moment, not a market's continuous verdict. The continuous verdict only begins at the IPO, which is precisely why one is being discussed.

The Run-Rate Trick: Why $4.6 Billion and $65 Billion Are Both True

This is the section to keep, because you'll need it for every AI headline from now until the IPO:

• 📐 A run-rate is a camera, not a film. Take the most recent month's (or quarter's) revenue and multiply it across a year: that's an "annualised run-rate". It answers "how big is this company RIGHT NOW, at this pace?", and for a business doubling every few months it will always dwarf the trailing year's actual total, because the trailing year includes all the smaller months.

• 🧮 So both Anthropic numbers are honest: calendar-2025 revenue was about $4.6 billion (the film of last year, per Reuters on the prospectus), while July 2026's pace annualises to about $65 billion (the camera, per Bloomberg). Neither is a lie. But a headline that swaps one for the other, or compares one company's run-rate against another's calendar year, is telling you a story the numbers don't.

• ⚠️ The two failure modes to watch for: run-rates assume the latest month CONTINUES, so they flatter anything growing fast and forgive anything about to slow; and "preliminary" or "adjusted" figures (like the reported quarterly profit) carry definitions you haven't seen. When the IPO prospectus becomes public, the audited film replaces the reported camera, and that's historically the moment hype meets accounting.

• 🧭 The habit: whenever an AI company's "revenue" makes headlines, ask three words: pace or past? It's the same anchor discipline as this week's Bumble lesson, a number means nothing until you know what it's measured against.

The IPO Watch: The Loudest Knock the Door Has Heard

• 📄 What's reported: a confidential IPO filing (reported in June, around the $965 billion round), investor discussions of a roughly $2 trillion listing valuation (FT, August; WSJ, September), speculation of a listing "as soon as October" (Forbes, in May), and Bloomberg reporting of a possible further raise. What's confirmed and scheduled: nothing. Confidential filings can sit for months and be withdrawn; treat every date as weather until a public prospectus lands. And the waiting room is famously crowded: SpaceX, whose Starlink arm already moves listed telecom markets when it stretches, has topped the someday-IPO list for years, an Anthropic listing would be jumping a very old queue.

• 🏛️ Why the scale would be historic: the largest IPO raises ever, Saudi Aramco's and Alibaba's, were in the $25-30 billion range. The reported discussions here would dwarf anything the stock market has listed, arriving in the same season GCash listed in Manila, Jio opened India's biggest-ever IPO, and, pointing the other way, Oura, the autumn's hottest IPO, pulled its listing despite four-times demand, the month the door swung hardest in years, in both directions. The Oura reminder matters most here: filed and listed are different stations, and even oversubscribed IPOs turn back.

• 🔎 What a public prospectus would finally show: audited revenue and losses (the film, not the camera), customer concentration, the real cost of the computing that powers every answer, and margins after those costs, the exact numbers today's $965 billion was negotiated without the public seeing.

• 🧭 The discipline between now and then: IPO anticipation is a story market. Prices, dates and "targets" will be floated, and none of them bind anyone. The only two events that change what a careful investor knows are a PUBLIC prospectus and a priced deal, everything before that is, in this week's language, spectacle before the earnings call.

How to Research the Story Today, Without a Single Private Share

You can't buy Anthropic. But the story already runs through listed companies you CAN research, three layers of them:

• 🏰 Layer one, the backers: Amazon (AMZN) has reportedly invested about $8 billion; Alphabet (GOOGL) holds a long-standing stake; 2026 reporting describes investment commitments from Nvidia (NVDA) and Microsoft (MSFT). When a private holding's valuation triples in a year, its listed owners' stakes mark up with it, quietly, inside their balance sheets. (And note the rhyme with this week's Jio post: Meta and Google's 2020 stakes there are the same pattern, giants holding private bets the market reprices later.)

• 🔧 Layer two, the picks and shovels: every Claude answer runs on advanced chips, which means the revenue ladder above flows, eventually, through TSMC, which just reported a record $23.6 billion quarter, and the machines that make those chips from ASML, which said this week that supply won't meet demand through 2028, one click away inside the TSMC piece. The AI labs compete; the toll booths collect from all of them.

• 🧺 Layer three, the index you may already own: Amazon, Alphabet, Microsoft and Nvidia are four of the S&P 500's largest members, so an ordinary index investor already holds the backers, and through them, slivers of the private story, with the swings diluted across 496 neighbours.

• 🎯 The research posture: these are EXPOSURE ROUTES to study, not a basket to buy, each backer is a giant with a thousand other stories, and the chip chain carries its own cycle, as this week's selloff showed. Examples to research, not recommendations.

The Honest Catch

• 🔍 Reported is not audited. Every revenue and profit figure above comes from journalism and company disclosures around funding rounds, serious sources, but not the audited, standardised statements public companies must file. Trackers already disagree at the edges (estimates of the July run-rate span several billions). The public prospectus, when it comes, is the first document that settles arguments.

• 🧮 The multiples assume the miracle continues. Fifteen-times-revenue pricing on a company this size has almost no precedent, it embeds years of continued hypergrowth AND a path to durable profit in an industry whose computing costs are enormous and whose price competition is intensifying. If growth merely becomes excellent instead of unprecedented, the maths changes violently.

• 🌪️ The macro can reprice everything, fast. This very week, a hot US inflation print knocked AI-linked stocks worldwide, chip giants fell on record results, and it wasn't even the season's first wobble: a single OpenAI headline gave AI stocks their worst day since August. Assets priced on far-future promises, which is what a 15x-revenue AI lab is, answer to interest rates like everything else, only more so.

• ⚔️ The rivalry cuts both ways: the reported revenue lead over OpenAI is preliminary, both companies are racing on models, pricing and compute deals, and leads in this industry have flipped before. Competitive snapshots age in months.

• 🎯 And for the eventual IPO, pre-register your discipline now: demand will be theatrical, allocation scarce, and day-one prices are a seller's chosen anchor, the Bumble lesson again, this time at the entrance. Decide what evidence (audited margins, customer concentration, cost curves) would justify what price BEFORE the roadshow decides for you.

What It Means for You

• 🧭 Translate "Anthropic earnings" correctly: there are none yet, there are reported run-rates, funding rounds, and a rumoured listing. Knowing the difference between reported and filed is this story's entire skill.

• 📐 Carry the run-rate question everywhere: pace or past? It will serve you for every AI headline between now and the prospectus.

• 🗓️ Watch two things only: a PUBLIC prospectus (the audited film at last) and any priced deal. Everything else, targets, timing chatter, leaked numbers, is weather.

How to Research This with Nemo.money

The Nemo.money app is built to help you research before you decide:

• Invest from Just $1: Fractional investing lets you explore the backers and the chip chain with small, research-sized amounts.
• Zero-Commission Trading: Buy and sell US-market stocks and ETFs without commission fees.
• AI-Powered Insights & Nemes: Explore data, sentiment and curated themed collections (Nemes) as a research starting point (for research, not recommendations).
• Earn 6% AER on Idle Cash: Uninvested cash in your wallet earns 6% AER, paid daily in USD while you wait for the prospectus to do the talking.

Frequently Asked Questions (FAQs)

Does Anthropic report earnings?

No. Anthropic is a private company: it files no public quarterly reports and holds no earnings calls. The widely quoted figures are reported revenue run-rates and funding-round disclosures from outlets such as Bloomberg, Reuters, the Wall Street Journal and CNBC, Reuters has reported calendar-2025 revenue of about $4.6 billion from the company's IPO prospectus, while the July 2026 annualised pace was reported around $65 billion.

Can I buy Anthropic stock?

Not today: Anthropic shares don't trade publicly. Reports describe a confidential IPO filing and investor discussions of a roughly $2 trillion listing valuation, but nothing is scheduled or guaranteed. Investors researching the story today look at reported stakeholders such as Amazon and Alphabet, reported committed investors such as Nvidia and Microsoft, and the chip supply chain, as research examples, not recommendations.

How is Anthropic worth $965 billion?

That was the valuation negotiated in May 2026's $65 billion Series H round, led by Altimeter, Dragoneer, Greenoaks and Sequoia, against a then-reported $47 billion revenue run-rate, roughly 15-20 times revenue depending on the figure used. A funding-round valuation is a negotiated price at one moment, not a continuous market verdict; the public market's verdict begins only if and when the company lists.

Is Anthropic bigger than OpenAI?

By reported valuation, yes since May 2026 ($965 billion versus OpenAI's reported valuation at the time), and Bloomberg/WSJ reporting put Anthropic's preliminary Q2 2026 revenue ($11.5 billion+) ahead of OpenAI's ($6.7 billion), with Anthropic reporting an adjusted operating profit against OpenAI's reported loss. These are preliminary, reported figures in a fast-moving rivalry, snapshots, not standings.

What is a revenue run-rate?

It annualises the most recent month's or quarter's revenue (multiplying it across a year) to show a company's current pace. For fast growers it will far exceed the trailing year's actual revenue, which is why Anthropic's reported $65 billion pace and its reported $4.6 billion calendar-2025 revenue are both true. Run-rates assume the latest pace continues, so treat them as a snapshot, not a promise.

Final Thoughts: The Camera and the Film

Somewhere between a reported $4.6 billion and a reported $65 billion sits the real Anthropic, and the honest answer is that nobody outside the company, not the trackers, not the headlines, not this article, can price it precisely until a public prospectus replaces reporting with audited accounts. That's not a criticism of the company; it's the nature of private markets, and the reason "Anthropic earnings" returns rumours instead of statements.

What an everyday investor CAN do is already clear: learn the run-rate trick so headlines can't blur pace and past, know the three exposure layers that exist today, backers, chip chain, index, and pre-decide what evidence the eventual prospectus must show before any roadshow adrenaline decides for you. The month the IPO door welcomed GCash, opened for Jio and closed behind Bumble may yet be remembered as the quiet period before its loudest knock ever.

Until then: pace or past, reported or filed, camera or film. The investor who keeps those pairs straight is ready for whatever gets listed next.

Explore global stocks and ETFs from $1 with zero commission on the Nemo.money app.

Nemo = Never Miss Out.

Stay informed. Stay ahead.

#Anthropic #AI #IPO #Claude #StockMarket #Investing #NemoMoney

Terms and conditions apply. This is not investment advice. Past performance is not indicative of future results. Your capital is at risk. See website for Risk Disclosure. Exinity ME Ltd (https://nemo.money) is regulated by ADGM's Financial Services Regulatory Authority.