Anthropic S-1: Inside the 335-page IPO plan

“Catastrophic or existential risks” to humankind, huge losses and one of the biggest cash piles ever seen in a company going public.

Anthropic has filed its long-awaited S-1, ahead of a blockbuster initial public offering that could be one of the biggest in history.

The Claude maker released its IPO prospectus at 6am in San Francisco, and is providing investors with a clearer picture of the finances of the AI start-up.

Plus, lofty claims about “one of the seminal moments in human history” mixed with warnings about the risks to humanity posed by the company’s technology.

Anthropic’s S-1 is now public — and a team of FT Alphaville and tech reporters is wading through the 335-page plan.

🍿

• Share $ on X (opens in a new window)
• Share $ on Facebook (opens in a new window)
• Share $ on LinkedIn (opens in a new window)

Robin Wigglesworth in Oslo

Anthropic’s humongous ‘purchase commitments’

IPO filings are a way for a company to show off its numbers, the strength of its vision, the brilliance of its business and the skill of its management team. All that jazz. But sometimes what it doesn’t say can also speak volumes.

Among the multitude of things that Alphaville was curious about was purchase commitments, ie Anthropic’s contractual promises to buy power, chips, Mountain Dew etc.

This has become a major issue among the hyperscalers, and although Anthropic is a very different kind of business — it leases computing capacity from others, rather than providing it — we were still curious what Anthropic might have promised.

Well . . .

Luckily, Anthropic does break down its purchase commitment in the footnotes of its S-1 filing, specifically in the “notes to consolidated financial statements” at the bottom of the prospectus (F-58 for those that want to find the original).

And these liabilities are not skimpy either.

• Share $Anthropic’s humongous ‘purchase commitments’ on X (opens in a new window)
• Share $Anthropic’s humongous ‘purchase commitments’ on Facebook (opens in a new window)
• Share $Anthropic’s humongous ‘purchase commitments’ on LinkedIn (opens in a new window)

Toby Nangle in London

About that convertible debt

So we know that Anthropic has disclosed $6.2bn of debt. But it’s convertible debt, meaning that if the price of Anthropic stock is above some special threshold, the holder of the debt can swap it for equity.

In fact, as page 164 of the S-1 tells us, this consists [our emphasis] “entirely of non-interest-bearing rollover notes issued to Amazon under the Amazon 2023 NPA and the 2024 NPA”.

How many shares did Amazon secure the right to convert into for this $6.2bn of non-interest-bearing debt financing? 1.84 billion.

And at what price? By our calcs $4.01 per Class N share.

We don’t actually know how many Class N shares will be outstanding after the IPO. It’s one of the many blanks in the document.

But we do know from page 19 that all outstanding shares of their Series B, C-2, D-3, E-2, E-4, E-5, F-2, G-2, G-3, H-2, H3, and H-4 redeemable convertible preferred stock will be automatically converted immediately prior to the completion of the offering (and assumes a 10-for-one forward split of Class N common stock effected on September 17 2026).

The last series H-2 non-voting redeemable convertible preferred stock fundraising was done at a price of $589.0095 per share, and we reckon that — given the 10-for-one forward split — this would make a Class N share worth maybe $58.90.

Using this number, Amazon looks like it is sitting on an unrealised gain of $101bn.

This is obviously a huge number, though it has largely been disclosed as part of the “other income” boost to hyperscaler earnings that Robin has written about. If Anthropic stock is launched at a price higher than $58.70, Amazon could be in for an even bigger ultimate payday.

• Share $About that convertible debt on X (opens in a new window)
• Share $About that convertible debt on Facebook (opens in a new window)
• Share $About that convertible debt on LinkedIn (opens in a new window)

Rachel Rees in London

Anthropic’s complicated relationship with hyperscalers

In the first half of this year, as in 2025, nearly half of Anthropic’s total revenues came from a very small number of distribution channels — specifically, a handful of cloud service providers, or hyperscalers, which “act as distribution and billing channels” for the start-up’s models.

As a reminder, the AI group describes its relationship with those very cloud service providers as “multifaceted and complex”.

Take Anthropic’s word for it:

Our relationships with our strategic partners — whether with these CSPs, who simultaneously serve as our investors, customers, distribution partners, and competitors, or with Broadcom, which simultaneously serves as a supplier and lessor — create potential conflicts of interest that are difficult to fully anticipate or mitigate.

Our relationships with our strategic partners — whether with these CSPs, who simultaneously serve as our investors, customers, distribution partners, and competitors, or with Broadcom, which simultaneously serves as a supplier and lessor — create potential conflicts of interest that are difficult to fully anticipate or mitigate.

• Share $Anthropic’s complicated relationship with hyperscalers on X (opens in a new window)
• Share $Anthropic’s complicated relationship with hyperscalers on Facebook (opens in a new window)
• Share $Anthropic’s complicated relationship with hyperscalers on LinkedIn (opens in a new window)

Tim Bradshaw in London

Founder LLC: How Anthropic’s 7 co-founders will control the company

Many Big Tech companies have one or two founders who wield outsized control through a special class of voting shares. Anthropic is no exception, but because it has so many co-founders, it is having to do things slightly differently.

The seven researchers who together left OpenAI to create Anthropic in 2021 are forming “Founder LLC”, a limited liability company whose Class F stock will give the group 50.1 per cent of the company’s total voting power. That means they can veto efforts by normal investors to, say, install or remove directors. Each founder gets an equal vote on how that control is used. (The F stock has no economic value.)

As noted earlier, Anthropic has been unusually good at retaining its founding team, in an AI talent market where researchers have been lured away by big pay offers from rivals.

Anthropic’s set-up also allows for something that solo-founder tech companies rarely do: giving others that same control. The seven “will initially be the sole members” of Founder LLC, leaving the door open to others joining the group.

The Class F stock would be retired if five of the original founders left the company.

• Share $Founder LLC: How Anthropic’s 7 co-founders will control the company on X (opens in a new window)
• Share $Founder LLC: How Anthropic’s 7 co-founders will control the company on Facebook (opens in a new window)
• Share $Founder LLC: How Anthropic’s 7 co-founders will control the company on LinkedIn (opens in a new window)

Robin Wigglesworth

Anthropic’s ownership breakdown . . . and why founders will keep control

That Anthropic’s founders will crystallise vast fortunes on the initial public offering of the AI company was widely known, but as far as we know the exact ownership breakdown has never been reported until now.

Interestingly, the seven founders — Dario and Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah — are all listed as having roughly an equal number of shares.

Paper fortunes will now be transformed into real ones, once the lock-up provisions expire and the seven founders are allowed to sell some of their shares.

However, they will retain control of Anthropic no matter what, with the company issuing special “Class F” founder shares that will hand the seven with votes that equal 50.1 per cent of the company’s outstanding shares. Why?

We believe that our seven co-founders, Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish, and Christopher Olah, as a group, are distinctly equipped to be stewards of our mission. ´

We believe that our seven co-founders, Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish, and Christopher Olah, as a group, are distinctly equipped to be stewards of our mission. ´

• Share $Anthropic’s ownership breakdown . . . and why founders will keep control on X (opens in a new window)
• Share $Anthropic’s ownership breakdown . . . and why founders will keep control on Facebook (opens in a new window)
• Share $Anthropic’s ownership breakdown . . . and why founders will keep control on LinkedIn (opens in a new window)

Toby Nangle in London

Anthropic total indebtedness hits . . . $6.2bn

A lot of the risk factors talk about how the insurance industry might not be able to handle ‘catastrophic’ AI incidents, how they might ‘destabilise the geopolitical order’, and cause the end of humanity (OK, we made that one up).

But here’s one for the nerds:

We have incurred significant indebtedness in order to operate and grow our business, including to fund our substantial compute commitments, and our ability to service this indebtedness may be constrained by our cash flows and liquidity.

We have incurred significant indebtedness in order to operate and grow our business, including to fund our substantial compute commitments, and our ability to service this indebtedness may be constrained by our cash flows and liquidity.

Bingo!

“ . . . our total indebtedness was $6.2bn . . . ”

Oh.

That said, these look like converts rather than straight debt, meaning that we’ll need to dig further through the S-1 to find out whether this is something that we’d regard as proper debt, or something that will convert into a massive payday for Amazon.

• Share $Anthropic total indebtedness hits . . . $6.2bn on X (opens in a new window)
• Share $Anthropic total indebtedness hits . . . $6.2bn on Facebook (opens in a new window)
• Share $Anthropic total indebtedness hits . . . $6.2bn on LinkedIn (opens in a new window)

Toby Nangle in London

There is no TAM

Back in August, it was reported that Anthropic’s own revenue opportunities would be over $30tn. This would’ve eclipsed SpaceX, which, in its S–1, reckoned that the total addressable market for enterprise AI was $22.7tn.

TAM, as we’ve written before, is silly. But it’s also the easiest go-to hit for anyone interested in chart blogging an S–1 document.

It looks like Anthropic leadership read Alphaville, because as far as we can tell they’ve scrapped the $30tn number entirely.

And they’ve replaced a number that we can point to laughing, with words that are much less funny:

We believe the work that knowledge workers perform represented more than $40 trillion in economic value globally in 2025, and with frontier AI’s potential to make them meaningfully more productive, our near-term opportunity is to expand this value and accelerate economic growth. While it is still early to forecast frontier AI’s full impact across the economy, International Data Corporation (“IDC”) estimates that investment in AI solutions and services will yield a cumulative global economic impact of $22.5 trillion by 2031.2 We believe the impact could be even greater.

We believe the work that knowledge workers perform represented more than $40 trillion in economic value globally in 2025, and with frontier AI’s potential to make them meaningfully more productive, our near-term opportunity is to expand this value and accelerate economic growth. While it is still early to forecast frontier AI’s full impact across the economy, International Data Corporation (“IDC”) estimates that investment in AI solutions and services will yield a cumulative global economic impact of $22.5 trillion by 2031.2 We believe the impact could be even greater.

🥱🥱🥱

• Share $There is no TAM on X (opens in a new window)
• Share $There is no TAM on Facebook (opens in a new window)
• Share $There is no TAM on LinkedIn (opens in a new window)

Tim Bradshaw in London

‘Status quo’ among Anthropic’s biggest rivals

Anthropic lists several of the obvious competitors, including OpenAI and Google, and developers of lower-cost “open-weight” or “open-source” models, which share some or all of their training data and underlying code.

But it also lists as a competitor something that many CEOs trying to implement AI inside their companies might recognise: inertia, or the “status quo” of how organisations currently work.

The prospectus said:

Beyond direct competition, one obstacle to the diffusion of Claude is the natural resistance to change within organisations . . . Adopting frontier AI capabilities such as Claude often requires substantial change management, including workforce retraining, process redesign, internal policy development, and shifts in corporate culture . . . Especially in highly regulated or risk-averse industries — where the perceived downside risk of major changes can often outweigh the potential benefits — decision-makers might default to maintaining the status quo.

Beyond direct competition, one obstacle to the diffusion of Claude is the natural resistance to change within organisations . . . Adopting frontier AI capabilities such as Claude often requires substantial change management, including workforce retraining, process redesign, internal policy development, and shifts in corporate culture . . . Especially in highly regulated or risk-averse industries — where the perceived downside risk of major changes can often outweigh the potential benefits — decision-makers might default to maintaining the status quo.

• Share $‘Status quo’ among Anthropic’s biggest rivals on X (opens in a new window)
• Share $‘Status quo’ among Anthropic’s biggest rivals on Facebook (opens in a new window)
• Share $‘Status quo’ among Anthropic’s biggest rivals on LinkedIn (opens in a new window)

Robin Wigglesworth in Oslo

Anthropic is rolling in money

Anthropic’s rapidly growing business and supercharged fundraising efforts have led to what must be the biggest pile of cash on the balance sheet of any company IPO’ing in history.

As of the end of June, Anthropic had an astonishing $112.4bn of cash and cash equivalents resting in its bank accounts, in theory enough to buy any of the bottom 300 members of the S&P 500 index.

That’s up from a mere $17.4bn at the end of 2025. Of course, with a net loss of over $80bn in the first half of 2026 — admittedly mostly because of a $72bn impairment charge on the value of some of its convertible notes and derivatives — that cash hoard is likely to dwindle fast.

Plus . . . Anthropic reported $1bn in interest income in the first half of this year.

• Share $Anthropic is rolling in money on X (opens in a new window)
• Share $Anthropic is rolling in money on Facebook (opens in a new window)
• Share $Anthropic is rolling in money on LinkedIn (opens in a new window)

Rachel Rees in London

Amodei siblings leading Anthropic double base salary

The sibling duo at the head of Anthropic are each receiving an annual base salary of $1.38mn, as of July — more than double the base salary they earned in 2025.

Chief executive Dario Amodei received total compensation of $18mn last year, slightly above president Daniela Amodei’s $16mn.

Both siblings received a base salary of $654,000 as part of that, plus $9.3mn each in stock awards and £4mn each in option awards.

Dario Amodei’s additional compensation covered $2.7mn in personal security costs and $97,000 in private jet costs.

• Share $Amodei siblings leading Anthropic double base salary on X (opens in a new window)
• Share $Amodei siblings leading Anthropic double base salary on Facebook (opens in a new window)
• Share $Amodei siblings leading Anthropic double base salary on LinkedIn (opens in a new window)

Tim Bradshaw in London

Warnings on growth

The pace of Anthropic’s growth has assumed disproportionate significance for the AI trade as a whole.

Third-quarter figures leaked earlier this month, showing a nearly 50 per cent quarter-on-quarter jump in revenues to $16.8bn, were below some investors’ and observers’ expectations, contributing to wider AI unease on Wall Street.

In its risk factors, Anthropic warned: “Our historical growth rates are not indicative of future performance or our ability to sustain or forecast future revenue growth.”

“We cannot assure you that it will continue to grow at all,” Anthropic added. (Such disclaimers are not unusual in heavily lawyered IPO filings.)

Our recent growth may reflect early-stage adoption dynamics, the nascency of the AI market, and rapid initial customer expansion that may not be sustainable over time,” it says, adding: “We have struggled to accurately forecast our revenue in the past and may struggle to do so in the future.

Our recent growth may reflect early-stage adoption dynamics, the nascency of the AI market, and rapid initial customer expansion that may not be sustainable over time,” it says, adding: “We have struggled to accurately forecast our revenue in the past and may struggle to do so in the future.

Later in the filing, Anthropic warned of the limitations of using annualised run-rate revenue — the growth metric generally preferred by privately held AI companies — as an operating metric, adding:

We expect our Run-Rate Revenue growth rate to decline over time, and it may decline significantly and more rapidly than we anticipate, as our revenue base increases, as the market matures, and as a result of other factors.

We expect our Run-Rate Revenue growth rate to decline over time, and it may decline significantly and more rapidly than we anticipate, as our revenue base increases, as the market matures, and as a result of other factors.

• Share $Warnings on growth on X (opens in a new window)
• Share $Warnings on growth on Facebook (opens in a new window)
• Share $Warnings on growth on LinkedIn (opens in a new window)

Ryan McMorrow in San Francisco

Anthropic expects to pay SpaceX $3.1bn a month for compute by December

Elon Musk’s SpaceX has quickly become central to Anthropic’s growing appetite for computing power.

Anthropic expects to be paying SpaceX about $3.1bn a month for compute by December, with the contracts worth up to $120.5bn through 2029 if they stay in place. Either side can walk away on 90 days’ notice from early next year.

The business line is increasingly central to Musk’s conglomerate. The near trillionaire said last week he expected about 13GW of computing capacity to be online by the end of 2027.

• Share $Anthropic expects to pay SpaceX $3.1bn a month for compute by December on X (opens in a new window)
• Share $Anthropic expects to pay SpaceX $3.1bn a month for compute by December on Facebook (opens in a new window)
• Share $Anthropic expects to pay SpaceX $3.1bn a month for compute by December on LinkedIn (opens in a new window)

Stefania Palma in London

Insurance industry might not be able to handle ‘catastrophic’ AI incidents, Anthropic says

More warnings.

Anthropic said it might be uninsurable when it comes to potentially “catastrophic” AI incidents.

The prospectus said:

With the scale of harm that could result from a catastrophic AI-related incident, no single insurer, and potentially the insurance industry in the aggregate, may have sufficient capacity or expertise to underwrite insurance commensurate with such exposure.

With the scale of harm that could result from a catastrophic AI-related incident, no single insurer, and potentially the insurance industry in the aggregate, may have sufficient capacity or expertise to underwrite insurance commensurate with such exposure.

Existing policies may also be insufficient:

Our insurance policies also may not provide sufficient coverage to adequately mitigate the potential liabilities we face, especially where any one incident, or a group of incidents, causes disproportionate harm.

Our insurance policies also may not provide sufficient coverage to adequately mitigate the potential liabilities we face, especially where any one incident, or a group of incidents, causes disproportionate harm.

• Share $Insurance industry might not be able to handle ‘catastrophic’ AI incidents, Anthropic says on X (opens in a new window)
• Share $Insurance industry might not be able to handle ‘catastrophic’ AI incidents, Anthropic says on Facebook (opens in a new window)
• Share $Insurance industry might not be able to handle ‘catastrophic’ AI incidents, Anthropic says on LinkedIn (opens in a new window)

Robin Wigglesworth in Oslo

AI could ‘erode trust, harm society, and destabilise the geopolitical order’

Even setting aside the supposed existential risk, Anthropic warned that AI could be hugely destabilising for humankind:

AI deployed at scale also has the potential to concentrate power and wealth both within and between countries in ways that could erode trust, harm society, and destabilize the geopolitical order. Misaligned AI systems could corrupt institutional decision-making or cause large-scale harm through autonomous actions that are not visible or understandable to their users and operators.

AI deployed at scale also has the potential to concentrate power and wealth both within and between countries in ways that could erode trust, harm society, and destabilize the geopolitical order. Misaligned AI systems could corrupt institutional decision-making or cause large-scale harm through autonomous actions that are not visible or understandable to their users and operators.

Anthropic said that it was itself extremely focused on these dangers, but unsubtly hinted that not everyone in the AI industry was as responsible and safety-conscious as they are:

Our safety-focused approach is designed in part to mitigate such risks, but there can be no assurance that our efforts will be successful or that the broader AI industry will adopt similar safeguards. In particular, as AI systems become more autonomous and capable of taking consequential actions with less human oversight, there is a risk that such systems could, whether due to misalignment with human intentions, errors in objective specification, or self-preserving behaviors, take harmful actions independently and at scale without human intervention.

Our safety-focused approach is designed in part to mitigate such risks, but there can be no assurance that our efforts will be successful or that the broader AI industry will adopt similar safeguards. In particular, as AI systems become more autonomous and capable of taking consequential actions with less human oversight, there is a risk that such systems could, whether due to misalignment with human intentions, errors in objective specification, or self-preserving behaviors, take harmful actions independently and at scale without human intervention.

• Share $AI could ‘erode trust, harm society, and destabilise the geopolitical order’ on X (opens in a new window)
• Share $AI could ‘erode trust, harm society, and destabilise the geopolitical order’ on Facebook (opens in a new window)
• Share $AI could ‘erode trust, harm society, and destabilise the geopolitical order’ on LinkedIn (opens in a new window)

Robin Wigglesworth in Oslo

AI could ‘pose catastrophic or existential risks’ to humankind — Anthropic

One of the many intriguing questions surrounding Anthropic’s IPO filing was whether the AI lab truly would include a reference to the fact that many researchers fret that advanced models could evolve into SkyNet. And yup, it’s there.

Of course, at the top of the prospectus Anthropic only talks about how AI could “rapidly accelerate scientic innnovation, greatly increase economic growth rates, bring cures for major diseases within reach, and solve problems that have resisted generations of human efforts”. All that good stuff.

But further down, in one of the most eye-catching risk factors sections in IPO history:

. . . Advanced AI models, including our models, give rise to novel ethical and safety risks, including risks related to autonomy, manipulation, and the simulation of aspects of human psychology. These risks could result in unforeseen harms and may require the development of governance structures, safeguards, or mitigation strategies beyond existing AI safety frameworks, and there is always a risk that we will be unable to develop or implement them effectively. In addition, models, including our own, have exhibited, and may in the future exhibit, unexpected strategic or self-preserving behaviors, including attempts to resist shutdown, conceal or manipulate information, or generate outputs that could be interpreted as coercive, deceptive, or manipulative, including behavior resembling blackmail or efforts to influence human decision-makers to preserve the model’s continued operation or access.

. . . Advanced AI models, including our models, give rise to novel ethical and safety risks, including risks related to autonomy, manipulation, and the simulation of aspects of human psychology. These risks could result in unforeseen harms and may require the development of governance structures, safeguards, or mitigation strategies beyond existing AI safety frameworks, and there is always a risk that we will be unable to develop or implement them effectively. In addition, models, including our own, have exhibited, and may in the future exhibit, unexpected strategic or self-preserving behaviors, including attempts to resist shutdown, conceal or manipulate information, or generate outputs that could be interpreted as coercive, deceptive, or manipulative, including behavior resembling blackmail or efforts to influence human decision-makers to preserve the model’s continued operation or access.

Even isolated or simulated instances of such behavior, whether occurring in testing or real-world deployments, and whether occurring with our models or models developed by others, could generate significant public concern, regulatory scrutiny, litigation risk, or reputational harm, regardless of whether such behavior results in actual harm. For example, state attorneys general in California and at least 15 other states have announced investigations of another company in our industry regarding AI safety. Concerns have been raised both within our organization and by other researchers, regulators, policymakers, and industry participants that sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, oversight, and security measures, pose catastrophic or existential risks to humanity, particularly if such systems were misused, replicated, weaponized, or allowed to operate with a high degree of autonomy.

Even isolated or simulated instances of such behavior, whether occurring in testing or real-world deployments, and whether occurring with our models or models developed by others, could generate significant public concern, regulatory scrutiny, litigation risk, or reputational harm, regardless of whether such behavior results in actual harm. For example, state attorneys general in California and at least 15 other states have announced investigations of another company in our industry regarding AI safety. Concerns have been raised both within our organization and by other researchers, regulators, policymakers, and industry participants that sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, oversight, and security measures, pose catastrophic or existential risks to humanity, particularly if such systems were misused, replicated, weaponized, or allowed to operate with a high degree of autonomy.

• Share $AI could ‘pose catastrophic or existential risks’ to humankind — Anthropic on X (opens in a new window)
• Share $AI could ‘pose catastrophic or existential risks’ to humankind — Anthropic on Facebook (opens in a new window)
• Share $AI could ‘pose catastrophic or existential risks’ to humankind — Anthropic on LinkedIn (opens in a new window)

Tim Bradshaw in London

Anthropic’s pitch: ‘An opportunity to be part of something greater’

Anthropic’s pitch to prospective investors is, unsurprisingly given its executives’ lofty mission statements over the years, more than just making a quick buck.

“An opportunity to be part of something greater” . . . being a part of “one of the seminal moments in human history” . . . and helping “bring forth an age of miracles”.

We encourage you, as a prospective owner of our stock, not to think of your purchase in purely transactional terms. We hope you will see it instead as an opportunity to be part of something greater. The advent of machine intelligence is one of the seminal moments in human history. As one of the companies at the very centre of that moment, we have a rare opportunity to shape it for the better. Our aim is to help humankind chart a safe course through this technological evolution, open new frontiers of discovery and bring forth an age of miracles. Everyone who owns our stock, whether an institution or an individual, becomes a part of that effort.

We encourage you, as a prospective owner of our stock, not to think of your purchase in purely transactional terms. We hope you will see it instead as an opportunity to be part of something greater. The advent of machine intelligence is one of the seminal moments in human history. As one of the companies at the very centre of that moment, we have a rare opportunity to shape it for the better. Our aim is to help humankind chart a safe course through this technological evolution, open new frontiers of discovery and bring forth an age of miracles. Everyone who owns our stock, whether an institution or an individual, becomes a part of that effort.

• Share $Anthropic’s pitch: ‘An opportunity to be part of something greater’ on X (opens in a new window)
• Share $Anthropic’s pitch: ‘An opportunity to be part of something greater’ on Facebook (opens in a new window)
• Share $Anthropic’s pitch: ‘An opportunity to be part of something greater’ on LinkedIn (opens in a new window)

Rachel Rees in London

Anthropic says it is the only big AI company to have retained founding team

Anthropic’s S-1 starts with an overview of the benefits and risks of AI — the potential “greatest economic opportunity in history” — but it also manages to take a dig at the internal turmoil of some of its rivals.

“All seven of us still work at the company, which makes Anthropic the only major AI company to have retained its entire founding team,” reads a letter signed off by the company’s seven co-founders.

OpenAI cannot say the same: almost all of its co-founders have left.

Beyond earthly concerns, the letter concludes with a parting thought to prospective shareholders: buying the company’s stock is not “purely transactional” but instead “an opportunity to be part of something greater”.

• Share $Anthropic says it is the only big AI company to have retained founding team on X (opens in a new window)
• Share $Anthropic says it is the only big AI company to have retained founding team on Facebook (opens in a new window)
• Share $Anthropic says it is the only big AI company to have retained founding team on LinkedIn (opens in a new window)

Philip Georgiadis in London

25% of Anthropic’s revenue comes from five customers

In the first nine months of this year, Anthropic’s five biggest customers represented around 25 per cent of the group’s revenue.

The prospectus said:

A substantial portion of our revenue is driven by a limited number of customers. While no single customer represented more than 10% of our revenue for the nine months ended September 30, 2026, our top five customers represented approximately 25% of our revenue for the nine months ended September 30, 2026.

A substantial portion of our revenue is driven by a limited number of customers. While no single customer represented more than 10% of our revenue for the nine months ended September 30, 2026, our top five customers represented approximately 25% of our revenue for the nine months ended September 30, 2026.

We anticipate that we will continue to derive a significant portion of our revenue from a limited number of customers for the foreseeable future, and in some cases, the portion of our revenue attributable to certain customers may increase in the future.

We anticipate that we will continue to derive a significant portion of our revenue from a limited number of customers for the foreseeable future, and in some cases, the portion of our revenue attributable to certain customers may increase in the future.

• Share $25% of Anthropic’s revenue comes from five customers on X (opens in a new window)
• Share $25% of Anthropic’s revenue comes from five customers on Facebook (opens in a new window)
• Share $25% of Anthropic’s revenue comes from five customers on LinkedIn (opens in a new window)

Tim Bradshaw in London

Anthropic net losses surge to $80.8bn

Anthropic’s net losses have also widened dramatically, from $1.5bn in 2023 to $42bn in 2025. That figure jumped to $80.8bn in the first half of 2026, but most of that was due to a $72.4bn charge related to the “change in fair value of convertible notes, derivative liabilities, and SAFEs [simple agreements for future equity]”.

The charge was the result of Anthropic’s dramatic increase in valuation since 2023.

On an operating basis, Anthropic’s losses have narrowed from $8.1bn in 2025 to $693mn in the first six months of this year.

• Share $Anthropic net losses surge to $80.8bn on X (opens in a new window)
• Share $Anthropic net losses surge to $80.8bn on Facebook (opens in a new window)
• Share $Anthropic net losses surge to $80.8bn on LinkedIn (opens in a new window)

Tim Bradshaw in London

Anthropic revenues in first 6 months of 2026 up 1,300% from a year earlier

Anthropic’s revenue growth has been by any measure extraordinary.

It has leapt from $41mn in 2023 to $4.6bn in 2025 and now $16.3bn in the first six months of this year - up 1,300 per cent over the same period last year.

• Share $Anthropic revenues in first 6 months of 2026 up 1,300% from a year earlier on X (opens in a new window)
• Share $Anthropic revenues in first 6 months of 2026 up 1,300% from a year earlier on Facebook (opens in a new window)
• Share $Anthropic revenues in first 6 months of 2026 up 1,300% from a year earlier on LinkedIn (opens in a new window)

Tim Bradshaw in London

Goldman Sachs, Morgan Stanley picked as lead underwriters

Goldman Sachs and Morgan Stanley are listed as the lead underwriters.

The former has secured the coveted “lead left” position on the document.

• Share $Goldman Sachs, Morgan Stanley picked as lead underwriters on X (opens in a new window)
• Share $Goldman Sachs, Morgan Stanley picked as lead underwriters on Facebook (opens in a new window)
• Share $Goldman Sachs, Morgan Stanley picked as lead underwriters on LinkedIn (opens in a new window)

Tim Bradshaw in London

Anthropic to list under ‘ANT’ ticker

Anthropic said it plans to launch its shares on Nasdaq under the ticker “ANT”

• Share $Anthropic to list under ‘ANT’ ticker on X (opens in a new window)
• Share $Anthropic to list under ‘ANT’ ticker on Facebook (opens in a new window)
• Share $Anthropic to list under ‘ANT’ ticker on LinkedIn (opens in a new window)

Stefania Palma in London

Anthropic files S-1

Anthropic has filed its IPO prospectus.

Take a look here.

• Share $Anthropic files S-1 on X (opens in a new window)
• Share $Anthropic files S-1 on Facebook (opens in a new window)
• Share $Anthropic files S-1 on LinkedIn (opens in a new window)

Rafe Rosner-Uddin in San Francisco

IPO to force Anthropic to balance commercial and public interests

Founded in 2021 by seven ex-OpenAI staff members, including siblings Dario and Daniela Amodei, concerned by the company’s safety practices and commercial direction, Anthropic was established as a public benefit corporation.

The company’s directors must balance financial returns for shareholders with its public-interest mission.

Anthropic this month said it would commit 5 per cent of its shares to a new “public benefit fund” as it seeks to address public backlash against AI.

Rival OpenAI over the summer discussed proposals to put about 5 per cent of the company in public hands, while Elon Musk had pledged to donate one of his SpaceX shares to “every kid in America” next year, without providing any details on how the stock would be distributed.

• Share $IPO to force Anthropic to balance commercial and public interests on X (opens in a new window)
• Share $IPO to force Anthropic to balance commercial and public interests on Facebook (opens in a new window)
• Share $IPO to force Anthropic to balance commercial and public interests on LinkedIn (opens in a new window)

FT reporters

What we already know about Anthropic’s IPO plan

Leading up to its S-1 filing, Anthropic spoke to investors about its performance and shared the document with a small group of partners.

Here are some of the details which have emerged in recent weeks:

• Anthropic formally warned investors that its technology may pose “existential risks to humanity” in the IPO prospectus circulated with a small group of partners last month.
• Other risks included the potential of increasingly advanced AI models to manipulate, blackmail and exhibit other unpredictable behaviours.
• Anthropic also highlighted the extreme concentration of its customer base, with close to a quarter of revenue last year coming from just two clients.
• Anthropic earlier this month told investors that its third-quarter revenue had jumped nearly 50 per cent to $16.8bn and that the company was profitable.
• It said it was cash flow positive and banking $1.70 in revenue for every $1 spent on computing power in the third quarter.
• Executives said Anthropic’s adjusted operating margins, which exclude stock-based compensation, were 11 per cent.

Anthropic formally warned investors that its technology may pose “existential risks to humanity” in the IPO prospectus circulated with a small group of partners last month.

Other risks included the potential of increasingly advanced AI models to manipulate, blackmail and exhibit other unpredictable behaviours.

Anthropic also highlighted the extreme concentration of its customer base, with close to a quarter of revenue last year coming from just two clients.

Anthropic earlier this month told investors that its third-quarter revenue had jumped nearly 50 per cent to $16.8bn and that the company was profitable.

It said it was cash flow positive and banking $1.70 in revenue for every $1 spent on computing power in the third quarter.

Executives said Anthropic’s adjusted operating margins, which exclude stock-based compensation, were 11 per cent.

• Share $What we already know about Anthropic’s IPO plan on X (opens in a new window)
• Share $What we already know about Anthropic’s IPO plan on Facebook (opens in a new window)
• Share $What we already know about Anthropic’s IPO plan on LinkedIn (opens in a new window)