Anthropic IPO: Claude Maker Files for Nasdaq Listing

Anthropic plans to list its Class A common stock on the Nasdaq Global Select Market under the ticker symbol ANT. Goldman Sachs is lead-left bookrunner, Morgan Stanley is alongside it, and the syndicate includes more than 20 other banks, according to coverage of the filing.

What Anthropic has not done yet is just as important as what it has done. It has not disclosed how many shares it intends to sell, or at what price. That means the most eye-catching numbers in the headlines — a valuation of $1.5 trillion to $2 trillion, and a raise of up to $100 billion — are reported estimates, not filed terms. They could move, and they probably will, once the roadshow and pricing process run their course. Reports put a roadshow starting around 9 November, with pricing the following week, and possible trading before the end of November, but none of those dates are final until Anthropic files an amended S-1 with a price range.

This article is not investment advice, and it isn't a recommendation to buy or avoid the shares. It's an explainer: what the filing actually says, what is still only reported, and what the listing could mean for people who use Claude every day.

The numbers that matter first

If you only remember five figures from the filing, make them these:

• First-half 2026 revenue: $16.3 billion, compared with about $1.2 billion in the same period a year earlier.
• First-half 2026 net loss: $80.7 billion, largely driven by a noncash accounting charge, according to reporting on the filing.
• Last private round: $65 billion raised in May 2026, at a $965 billion post-money valuation, according to Reuters.
• Reported IPO valuation range: $1.5 trillion to $2 trillion.
• Reported potential raise: up to $100 billion.

Put the revenue growth next to the loss, and you get the central tension of this IPO. Anthropic is growing extraordinarily fast in revenue terms — more than ten times year-on-year for the first half, on those figures — while also reporting an enormous accounting loss. That loss does not mean Anthropic spent $80.7 billion in cash in six months. A large part of it reflects fair-value accounting marks, which can swing violently with valuation changes. But it does mean the filing is not a simple "profitable company goes public" story. Anyone reading the S-1 needs to separate cash, operating profit, and accounting marks, because they tell very different stories.

How the $1.5 trillion to $2 trillion range works

Valuation ranges sound precise. They're not. They're a shorthand for what investors might pay, based on revenue, growth, and what comparable companies are worth.

Here's one way analysts in the coverage worked it out. Multiply Anthropic's third-quarter revenue by four, and you get a run-rate of about $67 billion a year. At a $1.5 trillion valuation, that's about 22 times run-rate sales. At $2 trillion, it's about 30 times. On trailing twelve-month revenue of roughly $36.5 billion, those multiples rise to about 41 times and 55 times.

For comparison, the same coverage put a reported OpenAI round at about $30 billion at a $1.25 trillion pre-money valuation, on about $50 billion of annualised revenue — about 25 times run-rate. Nvidia, on a similar basis, was put at about 13.5 times, and SpaceX at about 64 times. Those comparisons are useful, but they're also fragile: they depend on which revenue period you use, whether you annualise a single quarter, and how much growth you assume will continue.

The May private round gives another anchor. At $965 billion post-money, a $1.5 trillion IPO would mean public buyers paying about 55% more than private investors paid five months earlier. At $2 trillion, they'd pay about double. That's not impossible for a fast-growing AI company, but it does mean the IPO price has to be justified by growth that hasn't fully happened yet. As one venture investor quoted in coverage put it, investors should distinguish between what a company may ultimately be worth, and the valuation at which it should enter public markets. Those are two different questions.

SpaceX raised $75 billion in its June 2026 IPO, setting a record. If Anthropic raises $100 billion, as reported, it would surpass that record — but for users, what Anthropic does with the money matters more than the headline size.

The compute bill is the real story

Strip away the valuation talk, and the most important part of the S-1 may be the compute commitments.

Due-diligence coverage of the filing put named compute commitments at about $637 billion, against about $131.7 billion of cash. The filing also describes spending of up to $120.5 billion on computing linked to SpaceXAI and Nvidia, along with up to $114.5 billion with Google, through 2033. Those are not annual figures — they're commitments over years — but they show the scale of the bet Anthropic is making on access to chips, data centres, and cloud capacity.

That bet is unavoidable for a frontier AI lab. Training and running large models requires enormous, sustained compute, and that compute is now the main constraint on growth, ahead of talent or even customer demand. The question for investors is not whether Anthropic needs compute — it clearly does — but whether it can turn that compute into revenue fast enough, at high enough margins, to pay for the next round of commitments.

There's a related detail that matters for product pricing. Coverage of the filing notes that Opus pricing has fallen from $15 per million input tokens and $75 per million output tokens, to $4 and $20 respectively. That's a dramatic price cut, and it's part of a wider pattern across the industry: frontier models are getting cheaper to call, even as they get more capable. For developers, that's good news. For Anthropic's margins, it means volume and efficiency have to do a lot of work to offset falling prices per token.

Anthropic also disclosed that 5% of its shares will go to the Anthropic Public Benefit Fund. That's part of its structure as a public benefit corporation, and it's worth understanding before the IPO, because it affects governance in ways that go beyond a normal shareholder vote.

Governance: not a normal public company

Anthropic is not listing as a straightforward tech company with one share, one vote, and a board that answers only to shareholders. Its public benefit structure, and the governance arrangements described in S-1 coverage, include features that give long-term mission considerations a formal role.

Due-diligence coverage highlighted, among other catches, a single share carrying 50.1% of the vote, a Public Benefit Fund ratchet, and questions about transfers through special-purpose vehicles. The exact legal mechanics are complex, and they will be scrutinised heavily during the roadshow, but the direction is clear: Anthropic is trying to go public while keeping a structure that prevents short-term shareholder pressure from overriding its safety and benefit commitments.

For investors, it's a trade-off. You may be buying into a company whose governance deliberately limits how much control public shareholders have over certain decisions. That's not necessarily bad, but it's not the same as buying a normal listed stock.

What does this mean if you use Claude?

For most Claude users, the IPO filing changes very little in the short term — and that's an important point to make clearly.

A filing does not change Claude's pricing, its model lineup, or its availability overnight. It does not mean Claude is about to become free, or about to become dramatically more expensive, or about to be shut down. It means Anthropic is moving from private financing, with a small number of large investors and confidential disclosures, to public financing, with quarterly reporting, public scrutiny, and a much larger investor base.

Over time, that shift could matter in three ways.

First, disclosure. As a public company, Anthropic will have to report revenue, costs, and risks regularly. That will give users, researchers, and competitors a much clearer picture of how Claude is actually performing, where its costs are going, and where its strategy is heading. For a field that has often relied on selective announcements, that's a welcome change.

Second, capital. A large IPO raise, if it happens at the reported scale, would give Anthropic enormous resources to invest in compute, talent, and product development. That could mean faster model improvements, better infrastructure, and more ambitious products — including, potentially, more capable Claude models for coding, research, and agent work.

Third, pressure. Public markets reward growth and margins, quarter after quarter. That pressure can push a company to optimise for revenue in ways that don't always align with user interests — more aggressive upselling, tighter free tiers, or faster product changes. Anthropic's public benefit structure is designed to resist some of that pressure, but it won't eliminate it entirely. Watch how Claude's pricing and free access evolve in the months after listing, not just on listing day.

The risks, stated plainly

Every S-1 contains a risk section, and Anthropic's is no exception. Here are the themes that matter most, in plain language.

Revenue concentration and customer dependence: a large share of revenue comes through major cloud and distribution partners, including Amazon, Google and Microsoft, according to due-diligence coverage that put 48% of booked gross through those channels. That gives Anthropic reach, but it also means partner decisions can move its business quickly.

Compute dependence: Anthropic's future depends on access to chips and data centres at scale, at prices it can afford, over many years. If compute costs rise, or supply tightens, margins and growth suffer.

Competition: OpenAI, Google, Meta, and a growing set of open-weight labs are all competing for the same developers, enterprises, and users. Model quality gaps close quickly, and pricing pressure is intense, as the Opus price cut shows.

Valuation risk: at $1.5 trillion to $2 trillion, the market is paying for a lot of future growth. If growth slows, or if AI spending cools, a high multiple can fall fast, regardless of how good the underlying technology is.

Governance and regulatory risk: Anthropic's structure, its public benefit obligations, and the wider regulatory debate about AI safety, data, and market power could all affect how it operates as a public company. Washington, in particular, will be watching a listing of this size closely.

What happens next

If the reported timetable holds, the next steps are: a roadshow starting around 9 November, pricing the following week, and possible trading before the end of November. The key document to watch is not the next headline, but the amended S-1 that sets a share count and price range. That will turn a reported valuation range into a filed, testable number.

After that, watch three things: how the shares trade in the first weeks, what Anthropic says about its second-half 2026 results, and whether Claude's product and pricing change in ways that reflect its new public-company status.

Anthropic's IPO, if it completes at anywhere near the reported scale, will be one of the largest listings in history, and a landmark moment for AI as an industry. But its real significance won't be the valuation on day one. It will be whether a frontier AI lab can operate as a public company — with public disclosure, public pressure, and a public benefit mission — and still build models that people trust enough to use for their most important work.

That's the test Claude's maker has just set for itself. The filing is the start of it, not the end.