Anthropic S-1: What the IPO Filing Says About AI Spending

The Anthropic S-1 shows $33.2B in nine-month revenue and 11,900 big customers, as its IPO draws six times demand. What it means for AI budgets and marketers.

3D illustration of a lit glass block on a dark round plinth at the center of a mirror floor, circled by six rings of small glowing cyan and violet beads with a few warm amber beads among them.

In short: Read the Anthropic S-1 as a map of where enterprise AI spending is heading, not as a stock tip. Benchmark your own AI spend against the filing's customer tiers, expect a harder sales push as the company answers to public investors, and lock in pricing and data terms before your next renewal. Keep a second model tested so you can negotiate from strength.

On this page19

Overview

Demand for Anthropic's stock is running hot. Its IPO is about six times oversubscribed, Bloomberg reported on November 12, 2026, citing people familiar with the deal. A day earlier, it had said only "multiple times."

The bigger story for marketers is in the paperwork. Anthropic's amended S-1 filing gives the clearest public view yet of how much companies spend on AI, and how fast that spend is growing.

We covered the early valuation talk in our first Anthropic IPO guide. This one goes through the filing itself: the deal terms, the customer numbers and what they mean for marketing budgets.

Disclosure: Vanaxity builds AI agents for marketing and works with several AI model providers. This article was drafted with help from Claude. It isn't investment advice. Its facts come from Anthropic's SEC filing and from Bloomberg, Quartz and other press reports published between November 10 and 13, and we label our own math wherever we use it.

Key takeaways

  • Anthropic is selling 825 million shares at $70 to $78 each, to trade on Nasdaq as ANT.
  • Bloomberg: about six times demand.
  • Seven tech companies will buy $26.0 billion more at the IPO price in a private placement.
  • Revenue hit $33.2 billion in the first nine months of 2026, up from $2.5 billion a year earlier.
  • More than 11,900 customers spent $100,000 or more over the past year.
  • Fortune 500 customers still spend less than 0.1% of their total expenses with Anthropic, so expect a big sales push.

Book a free fit checkMap your SEO, GEO and AEO workflow before you build.

Book a consultation

The Anthropic IPO terms at a glance

Here's what the November 10 filing and reports say about the deal:

TermDetail
Shares offered825,000,000 Class A shares
Who sells801.4 million new shares from Anthropic, 23.6 million from existing holders
Price range$70 to $78 per share
Exchange and tickerNasdaq Global Select Market, ANT
Private placement$26.0 billion at the IPO price
Retail shareUp to 20% of IPO shares
Expected pricingNovember 18, with trading from November 19, per reports

The private placement buyers are AMD, Broadcom, Google, Micron, Microsoft, Nvidia and TSMC Partners.

At the top of the range, the IPO raises about $64.4 billion, by our math. Add the placement, and the total is about $90.4 billion. Bloomberg says that would make it the largest listing ever. Bigger than SpaceX.

Reports put the fully diluted value at about $1.4 trillion, per Quartz. That's below the $1.5 trillion figure some investors floated earlier this month, which we covered in our first guide.

Who is buying the $26 billion private placement

The private placement is unusual in size and in who's buying. Seven companies agreed to buy $26.0 billion of a separate share class at the IPO price, once the IPO closes.

Look at the list: AMD, Broadcom, Google, Micron, Microsoft, Nvidia and TSMC Partners. These are the firms that make Anthropic's chips, host its models or sell them to customers. That's no accident.

Why does that matter to buyers? It ties Anthropic closer to your clouds. That can make Claude easier to buy through existing contracts. It can also make it harder to tell where one vendor's interests end and another's begin.

The takeaway? If you buy Claude through a cloud marketplace, ask which discounts and credits stack, and which don't.

Anthropic IPO oversubscribed: what six times means

Oversubscribed means buyers asked for more shares than exist. Six orders per share.

That doesn't mean everyone who asked will get a slice, or even close to what they ordered, once the books close. Underwriters told buyers Anthropic plans to favor its current backers, Bloomberg reported. So many new investors will get far less than they asked for.

Retail investors get an unusual share. The filing says up to 20% of IPO shares could go to individuals, through brokers such as Schwab, Fidelity and Robinhood. Some coverage calls that about double the usual level for a big US listing, a sign of how much Anthropic wants everyday Claude users to own a piece, per InvestingLive.

Strong demand also helps the price. When orders pile up, companies often price at or above the top of their range. That would raise the totals above.

What the Anthropic S-1 shows about revenue

The revenue line is the headline. Here's how it has grown, per the filing:

PeriodRevenueGrowth
2023$41 millionNot given
2024$386 million841%
2025$4.6 billion1,088%
Nine months to September 30, 2026$33.2 billion1,206% vs a year earlier
Third quarter of 2026$17.0 billionUp from $1.4 billion

That puts trailing 12-month revenue at about $35.3 billion as of September 30, by our math. We added the $33.2 billion from the first nine months to roughly $2.1 billion from the last quarter of 2025.

The business also lost less money. Operating loss fell to $1.3 billion in the first nine months of 2026, from $5.2 billion a year earlier. Anthropic says revenue per employee rose to over $6.6 million.

The filing is blunt about the future, though. Anthropic says it doesn't expect growth to continue at these rates and can't promise it will grow at all.

How to read the Anthropic S-1 in 15 minutes

An S-1 runs hundreds of pages, and this one is long even by those standards. You don't need to read it all. Four parts give you most of what matters for a vendor review:

  • The prospectus summary, for the business in a few pages.
  • Key metrics, for customer counts and revenue by tier, which is where the enterprise AI spending numbers in this guide come from.
  • Risk factors.
  • Management's discussion, for why the numbers moved.

Skim the summary first, then jump to the metrics. Read risk factors with your own contract in mind. The goal isn't to judge the stock. It's to understand the company you rely on.

11,900 big customers: what the Anthropic S-1 says about AI spending

For marketers, the customer numbers matter more than the stock price. They show how many companies now spend real money on AI.

Anthropic customers spending $100,000 or more a year

Customers, trailing 12 months

  1. December 202320
  2. December 2024200
  3. December 20251,500
  4. September 202611,900
From about 20 at the end of 2023 to about 200, about 1,500 and more than 11,900 by September 30, 2026, per Anthropic's S-1. Counts are approximate in the filing.Source: Anthropic S-1/A, November 10, 2026

So the count went from about 20 at the end of 2023, to about 200, to about 1,500, to more than 11,900 by September 30, 2026. Those customers brought in $24.5 billion in the first nine months of 2026, or about 74% of revenue.

The top tiers are growing too. More than 1,900 customers spent $1 million or more, and about 250 spent $10 million or more. Over 50% of the Fortune 500 now count as $100,000-plus customers.

Enterprise AI spending benchmarks from the Anthropic S-1

Few vendors publish numbers like these, so they make a useful benchmark for enterprise AI spending. Here's the set in one place, as of September 30, 2026:

MeasureFigure
Customers at $100,000 or more a yearMore than 11,900
Customers at $1 million or moreMore than 1,900
Customers at $10 million or moreAbout 250
Share of revenue from $100,000-plus customersAbout 74%
Fortune 500 companies at $100,000 or moreOver 50%
Global 2000 companies at $100,000 or moreOver 30%

Use these as rough guides, not targets. Your right level of spend depends on your work, not on what other companies pay.

AI is still a tiny share of big budgets

Here's the line every marketing leader should read. Anthropic says its Fortune 500 customers spent, on average, less than 0.1% of their total expenses with it in the third quarter of 2026.

Its top 10 Fortune 500 customers spent less than 1%. Anthropic calls that more than eight times the average, and says it shows a lot of room to grow.

Read that from the buyer's side. Your AI vendor sees your company as early in its spending. As a public company, it will be under pressure to grow that share every quarter.

That's not bad news. It means more products, more pilots and more offers. But it also means a sharper sales push. Buyers who know their numbers will get better deals.

How Anthropic sells: lessons for B2B marketers

The S-1 also explains how Anthropic grew so fast without a giant sales team. It credits what it calls intelligence-led growth and broad distribution partnerships.

In plain words, the product did much of the selling, and partners did much of the distribution. Claude reaches buyers through Amazon, Google and Microsoft clouds, as well as directly.

Three lessons stand out:

  • Let usage sell. A product that proves its value in a short pilot needs far less persuasion when the time comes to sign a bigger deal.
  • Use partner channels. Being in a buyer's existing cloud or marketplace removes friction.
  • Track expansion, not just new logos. About 74% of Anthropic's revenue comes from customers spending $100,000 or more a year.

None of this is new advice. But it's rare to see it confirmed at this scale, with audited numbers behind it.

What the Anthropic S-1 means for marketing teams

If your team uses Claude, directly or inside other tools, here's what to expect after the listing:

  • More sales outreach, with offers tied to volume and longer contracts.
  • New products aimed at knowledge work, a group the filing expects to top one billion workers by 2028.
  • Pricing that public investors will watch closely, quarter by quarter.
  • More disclosure, which makes vendor risk easier to judge.

Our earlier guide covers how to map where your team relies on Claude and how to keep a second model ready. Those steps matter more now, not less. See the first Anthropic IPO guide for the full checklist.

The IPO as a brand moment

There's a marketing story in the deal structure itself. Setting aside up to 20% of shares for individuals, through apps like Robinhood and brokers like Schwab and Fidelity, turns users into owners.

That builds a kind of loyalty money can't easily buy. People who own a slice of a company tend to follow its news, defend its products and try its launches.

Most brands can't run an IPO for buzz. But the idea scales down. Give your best customers a real stake in what you build, such as early access, a say in the roadmap or a visible role in your community, and they'll act more like partners than buyers.

A worked example with our own math

Here's how to use the filing to benchmark your own AI spend, with made-up numbers rather than data from any real company, so treat the result as a method, not a finding.

Say your company has $500 million in yearly expenses. The filing says Fortune 500 customers average under 0.1% of their total expenses with Anthropic, so let's use 0.1% as a generous upper mark for a company your size. At 0.1%, that would be $500,000 a year.

Now say your team spends $120,000 a year on Claude across tools. That's 0.024% of expenses, well under the average. It also puts you in the $100,000-plus tier the S-1 counts as an enterprise customer.

What does that tell you? You're a customer the vendor wants to grow. That gives you room to ask for volume pricing, better data terms or pilot credits before you commit more. Our guide to FinOps for AI agents shows how to track that spend by team and tool.

What the IPO doesn't change

It's easy to overreact to a listing this big. So what stays the same?

The models don't change because the company is public. Your API keys, your workflows and your current contract terms stay in place. Anthropic stays a public benefit corporation, with the same stated mission.

Claude also stays available through the same channels, from Anthropic directly and through the big cloud marketplaces. If something does change, it will come through product and pricing updates, not through the IPO itself. Watch those, not the ticker.

Questions to ask your AI vendor after the listing

A public listing is a good moment to check your terms. Here are five fair questions for any AI model provider, not just Anthropic:

  • Will our current price hold through the contract term?
  • What volume tiers or commitments unlock better rates?
  • How is our data used, stored and kept out of model training, and can we get that promise written into the contract rather than a policy page?
  • What costs extra?
  • What notice do we get before a model we use is retired?

Write the answers into your contract, not just your notes. Sales promises change faster than signed terms.

Risks the Anthropic S-1 spells out

Every S-1 lists risks, and this one has plenty. A few matter for buyers, not just investors:

  • Growth may slow sharply, as the company itself warns.
  • Compute costs are huge, and depend on a few chip and cloud partners.
  • Anthropic is a public benefit corporation with a novel governance setup, so it may weigh safety goals against profit.
  • Lock-up periods end later, which can move the stock and the mood around the company.

None of these changes how Claude works for you today. But they're worth knowing when you sign a multi-year deal.

What to watch next

Key dates and signals for the coming weeks:

  • Final pricing, expected on November 18
  • First day of trading as ANT, expected on November 19
  • Whether the price lands above the $70 to $78 range
  • Anthropic's first earnings report as a public company
  • Any changes to Claude pricing or enterprise terms after the listing

We'll update this guide after the pricing. For more on AI vendors and marketing, see our insights, or how we help brands build AI workflows that don't depend on one model.

Anthropic S-1 FAQ

When does Anthropic start trading?

Reports say Anthropic plans to price its IPO on November 18, 2026, and start trading on Nasdaq under the ticker ANT on November 19. Dates can change until pricing is final.

How many shares is Anthropic selling, and at what price?

The amended S-1 of November 10, 2026 offers 825 million Class A shares at $70 to $78 each. Anthropic sells about 801.4 million, and existing holders sell about 23.6 million.

How oversubscribed is the Anthropic IPO?

Bloomberg reported on November 12, 2026 that orders were about six times the shares on offer, citing people familiar with the deal. Underwriters reportedly plan to favor existing backers.

What is Anthropic's revenue?

The S-1 shows revenue of $33.2 billion for the nine months to September 30, 2026, up from $2.5 billion a year earlier, and $4.6 billion for 2025. By our math, that's about $35.3 billion over the trailing 12 months.

How many big customers does Anthropic have?

More than 11,900 customers spent $100,000 or more over the 12 months to September 30, 2026, up from more than 950 a year earlier. More than 1,900 spent $1 million or more, and about 250 spent $10 million or more.

Tran Tien VanFounder, Van Data Team. Runs Vanaxity, the done-for-you SEO, GEO and AEO content service for B2B companies.Researched and first drafted with AI assistance. Facts, figures and quotes link to their sources in the text.Connect on LinkedInHow we write Insights

Keep reading

View all
3D illustration of a lit glass tower on a dark mirror floor, with a faint neon outline above it where a taller top used to be and a ring of amber light where the tower now stops.
Anthropic IPO
Governance & ROIAnthropic IPOAnthropic valuation
Governance & ROI · Nov 2, 2026

Anthropic IPO: What the $1.5 Trillion Talk Means for You

Anthropic, the maker of Claude, filed its S-1 on October 26, 2026 to list on Nasdaq under the ticker ANT. No price range or date is set, but reports point to a roadshow from about the week of November 9 and a debut before Thanksgiving on November 26. Reports say it could raise $100 billion or more. The valuation talk has slipped: after an investor day on October 14, Reuters reported that some investors discussed about $1.5 trillion, down from roughly $2 trillion, because third-quarter revenue grew 45% to $16.8 billion after more than doubling in the second quarter. On October 30 the Financial Times reported that Blackstone, an anchor investor in a $63 billion chip financing, now expects Anthropic to get a double-B credit rating, the top tier of junk, after the IPO. No rating agency has issued that rating yet. The S-1 shows revenue of about $4.6 billion in 2025 and $16.3 billion in the first half of 2026, with a net loss of $80.7 billion that is mostly a $72.4 billion accounting charge. For marketers, the IPO matters less as a stock and more as a supplier story: about 80% of Anthropic's revenue is usage-based, it has warned of tight computing capacity, and a public company faces pressure to grow margins.

A central obsidian tower routing gold and cyan circuits to dozens of glass agent nodes across a dark platform.
Autonomous Marketing Agent
Governance & ROIThe scaling and institutionalization of autonomous marketing agents: governance as the enterprise unlockreasoning and acting agents
Governance & ROI · Jul 6, 2026

The Scaling and Institutionalization of Autonomous Marketing Agents: Governance as the Enterprise Unlock

The scaling and institutionalization of autonomous marketing agents is the shift from one-off prompt experiments to agentic AI that runs marketing work as a permanent operating function - and governance, not raw model capability, is now the enterprise unlock.

AI Agent ROI: How to Measure Cost per Successful Task
AI Agent ROI: How to Measure
Governance & ROIAI agent ROIuseful intelligence per dollar
Governance & ROI · Jul 18, 2026

AI Agent ROI: How to Measure Cost per Successful Task

Your automated SDR can send messages, AI copywriter produce drafts, and media-buying agent run continuously—yet token bills and seat licenses miss how much quality-approved work actually succeeded, leaving real value invisible.

Google ATLAS AI Study: Collaborator, Not Automator
Google ATLAS AI Study
Governance & ROIGoogle ATLAS AI studyreasoning and acting agents
Governance & ROI · Jul 23, 2026

Google ATLAS AI Study: Collaborator, Not Automator

B2B marketing, RevOps, and content teams evaluating autonomous agents face a specific decision: whether to redesign work for hands-off automation when Google’s ATLAS data shows workplace AI is used mainly as a collaborator.

OpenAI Rule-Based Rewards for Brand-Safety Governance
OpenAI Rule-Based Rewards
Governance & ROIOpenAI Rule-Based Rewardsrule-based rewards for language model safety
Governance & ROI · Jul 24, 2026

OpenAI Rule-Based Rewards for Brand-Safety Governance

Marketing teams adopting OpenAI Rule-Based Rewards as a governance model risk confusing a 2024 model-training technique for safety behavior with a tool that checks brand compliance or factual accuracy.

Microsoft Path to Production for Agents: Governance Playbook
Microsoft Path to Production
Governance & ROIMicrosoft Path to production for agentsMicrosoft Azure AI Tech Accelerator
Governance & ROI · Jul 27, 2026

Microsoft Path to Production for Agents: Governance Playbook

A checklist for moving agents from prototype to production, with permissions, logs, evaluation, and incident handling.