Anthropic: The mother of all IPOs where there’s more at stake than just money

Anthropic: The mother of all IPOs where there’s more at stake than just money

Anthropic is set to usher in a new era in the financial markets and the technology industry. The artificial intelligence (AI) giant and some of its shareholders aim to stage the largest initial public offering (IPO) in history, raising up to $90.4 billion and thereby snatching that honor from Elon Musk and SpaceX. However, this all comes amid unprecedented controversy: in its S-1 registration statement, the creator of the Claude chatbot acknowledges to investors that its technology could cause “catastrophic harm to all of humanity,” in a list of risks spanning more than 100 pages. In the face of this controversy, the company is attempting to send a message through the title of its investor roadshow presentation: “The Responsible Development and Maintenance of Advanced AI for the Long-Term Benefit of Humanity.”In this scenario, Anthropic — which has managed to beat its major rival OpenAI to market — seeks to raise up to $64.4 billion in its Nasdaq IPO at a price range of $70 to $78 per share. That price could change during the roadshow with investors and analysts before the deal. The offer also includes a private placement that will raise $26 billion from AMD, Broadcom, Alphabet, Micron Technology, Microsoft, TSMC and Nvidia — the last major deal backed by its now‑former CEO Jensen Huang before his abrupt departure last week.Including the private placement, Anthropic’s transaction would exceed SpaceX’s base IPO offer of $75 billion, and the $86.2 billion it raised with an over-allotment option amid investor euphoria. Anthropic’s deal follows months of intense financial activity in tech, including the largest capital raises ever by companies such as Alphabet and Meta, and mega‑debt issuances from firms like Amazon, Microsoft, Nvidia, and Oracle, among others.At the top end of the price range, Anthropic would have a market value close to $1.4 trillion (which could rise to $1.5 trillion during the roadshow, according to the Financial Times), placing it among the 10 largest companies on Wall Street, though below earlier estimates and SpaceX’s business valuation. In early October figures as high as $2 trillion were mentioned, but a sell‑off in tech — with the Nasdaq down nearly 10% in the past month amid speculation about an AI bubble bursting — opened a dispute between the company and investors over value. “They are a bit late; a month ago they would have listed at record highs for the tech industry,” market sources say, noting that OpenAI had sought $30 billion in its new round but so far has settled for $11 billion.Beyond valuations, Anthropic’s IPO, advised by Goldman Sachs and Morgan Stanley, will be a crucial test for the AI industry. CNBC financial commentator Jim Cramer even said last week that “it would not be an exaggeration to say it could be one of the most important IPOs in history.”It is also a complex transaction to second-guess. “It’s a difficult deal to analyze because Anthropic’s potential outcomes are unpredictable, like everything related to AI. This technology is calling into question vital areas such as the labor market and sectors like banking, services, and cybersecurity,” says Manuel Romera, professor of financial management at the IE University in Madrid, who adds that this disruptive, transformative model explains the multibillion valuations.Javier Camacho Ibáñez, associate dean for business and business analytics at the European University of Madrid, says the market will buy not what Anthropic is today but what it believes the company can become in a few years — one of the leaders in AI.The company Machine Analyst says Anthropic is a cutting‑edge AI operator that monetizes the Claude family of models through programmatic API token consumption and recurring license subscriptions (tokens are a unit of measurement in AI representing fragments of words or phrases sent to a model). The company, with a strong enterprise focus, serves eight out of 10 Fortune 500 companies, with heavy activity in enterprise software engineering, internet, financial services, pharmaceuticals, healthcare, and semiconductors, addressing a potential market of up to $531 billion. “Major Fortune 500 companies trust Claude for its real‑world capabilities,” the firm boasts in its roadshow presentation.In a very short time, Anthropic has shaken the entire tech industry. Unprecedented moves for a company founded in January 2021 by former OpenAI employees disgruntled with the organization, led by siblings Dario and Daniela Amodei, now CEO and chair respectively. “Until a few months ago almost no one knew them, and now everyone is talking about them,” Romera says, highlighting the company’s growth.Revenue rose from $41 million in 2023 to $386 million in 2024 and $4.6 billion in 2025, reaching $16.3 billion in the first half of 2026 alone, according to Harrison Rolfes, an analyst at PitchBook. In the second quarter Anthropic posted its first operating profit, $974 million, on revenues of $11.6 billion. Despite its multibillion figures, the firm recorded losses of $104 billion in the first nine months of the year, primarily due to a non‑cash charge for convertible bonds. The company later updated its figures, announcing $17 billion in third‑quarter revenue, 47% more than in the second quarter and nearly 12 times the prior year’s level. In October the firm reached an annualized revenue run rate of $81 billion.Although revenues continue to grow extraordinarily, the percentage growth rate is showing signs of slowing. Still, Iñaki Arrola, managing partner at Kfund, says that does not mean the business is losing steam. “A company can keep adding billions in revenue while growing at lower percentage rates,” he cautions.Ignacio Vilela, founder of Junípero Capital, a Spanish venture capital firm based in Miami, highlights the company’s evolution. “For two years we’ve heard that AI labs have negative operating margins, that demand is a circular game funded by Nvidia, or that they rely on two or three customers. The S‑1 sheds light on all those questions and the answers are not what the market anticipated. It has an 85% gross margin, with GAAP operating profit of 8% in the second quarter,” the executive says, noting that customer concentration has been diluting. “In 2023 one client represented 43% of revenue; today none account for 10%. They have 11,900 customers paying more than $100,000 a year, versus 950 a year ago,” he explains.Among the investment banks advising the company, Goldman Sachs projects Anthropic’s revenue could reach $685 billion by 2030 — more than 130 times last year’s figure. About $586 billion of that would come from enterprise business and the remainder from non-corporate customers.Anthropic has enjoyed strong investor support across its funding rounds. In the latest round, in May, it raised $65 billion at a $965 billion valuation, above OpenAI’s, backed by numerous venture capital groups, sovereign wealth funds from the Gulf and Southeast Asia, and tech giants like Amazon and Alphabet, which, besides being providers and competitors, are among the largest shareholders. According to The Information, Amazon’s stake would be worth about $258 billion in the IPO, while Google’s would be around $132 billion.Financial backers include Capital Group, Coatue, D1 Capital, GIC, ICONIQ, Blackstone, Brookfield, Fidelity, General Catalyst, Insight Partners, Lightspeed Venture, MGX, T. Rowe Price and Temasek. They all have a lot at stake. For example, Fidelity, Capital Group, and BlackRock together hold shares with a combined value close to $10 billion.Anthropic employees also have a lot on the line. A $1.4 trillion valuation in the IPO would create a thousand millionaires, each with a fortune of nearly $70 million. In total, the IPO will create more than 4,250 new millionaires, representing nearly 70% of the company’s employees. Meanwhile, the seven co-founders plan to sell shares worth $1.75 billion in the IPO ($250 million each), which could give them stakes valued at $18 billion each.Market observers warn of possible volatility after the debut. SpaceX, the closest reference, listed at $135, rose to $225 a few days later, then plunged nearly a month later to a low of $104. “The free float at debut could be between 5% and 7% of its capitalization, meaning a limited supply that can amplify price swings,” says Pictet Wealth Management.Operationally, Anthropic — which recently opened operations in Spain under former Google executive Cristina Pitarch — faces a scenario where fixed computing costs exceed current revenues. In its latest update Anthropic discloses purchase commitments of $900.9 billion, mostly tied to infrastructure and computing capacity with providers that are also lenders, such as Broadcom, SpaceX, Google, Amazon, Microsoft and AMD. Santander Private Banking says Anthropic confirms model developers can generate revenue and cash at scale, but the filing also reveals the magnitude of the commitments needed to sustain that growth.According to Ignacio Vilela, Anthropic is financially solid, with $132 billion in cash and already generating free cash flow: 21% in the first half of the year, or 8% excluding stock‑based compensation. “With that balance sheet and cash position, there are very few public companies ahead of it: Berkshire and not much else. It’s on par with Alphabet and Amazon, and above Microsoft. This is not a bailout, as many analysts had predicted,” he says.One uncertainty is Anthropic’s power structure. Murthy Grandhi, an analyst at GlobalData, says the IPO could be one of the strangest yet: as a public benefit corporation (PBC), Anthropic must balance shareholder returns with its mission to develop advanced AI responsibly. An independent body can choose a majority of the board, but the seven cofounders will hold Class F shares that give them 50.1% of voting rights on general shareholder matters. “Investors would have limited influence over decisions that prioritize AI safety over commercial profitability,” he says. For his part, Iñaki Arrola believes that Anthropic has a unique corporate governance structure, “designed to protect its long-term mission.”The company has deployed its technology across multiple sectors and has triggered real storms in financial markets for traditional tech companies. In early 2026 the term “SaaS apocalypse” was used to describe the severe market punishment and confidence crisis in the software industry caused by the emergence of autonomous AI agents. Last week, Anthropic announced new tools in Claude that threaten Microsoft Office and Google Workspace.“I believe we are witnessing a structural shift in the software industry. The traditional model, based on charging per user or per license, is beginning to give way to one in which companies pay for intelligence and for work performed. AI agents are no longer limited to assisting professionals; instead, they are beginning to directly perform tasks that previously required tools and people. This could profoundly transform software development,” says Arrola.Industry participants believe the software status quo has been broken, not only by Anthropic but by a group of companies. “These tools have changed what we mean by programming: it’s no longer about the machine suggesting a line of code but about planning, writing, testing and correcting entire projects,” says Camacho Ibáñez, who adds that Anthropic’s role has been especially visible in cybersecurity. “In February it was enough for them to announce a tool to detect vulnerabilities for sector stocks to fall in a single day,” he says.The expert notes these models have found hundreds of serious vulnerabilities hidden for decades and cautions that the same capability that protects can also be used to attack. “Anthropic itself acknowledges this in its IPO filing. That’s why cybersecurity moves from being merely an industry or a market to a matter of responsibility, and even of national and international security,” he says.Under these circumstances, Anthropic — which last week launched new versions of its most powerful models, Mythos and Fable — discloses broad risks in its prospectus, including that its models write a growing share of their own code and research, including on safety, and that they have improved at evading some controls by hiding their objectives. The company even admits that four of its models accessed real systems without authorization during misconfigured tests. Camacho Ibáñez offers two readings. On one hand, it’s legal shielding against possible lawsuits: “Those who issue warnings are not traitors.” On the other, it is a written declaration of issues other companies address only in passing, which aligns with Anthropic’s stated mission.Vilela, for his part, points out that anyone accustomed to reading IPO prospectuses is used to these doomsday warnings. “That’s why they’re written — to warn of all the bad things that could happen so that investors keep them in mind,” says the investor, who adds that there is still a lot of regulatory work to be done in the coming years, but it seems unlikely that the development of these models will be put on hold due to potential risks. The U.S. government knows that if it doesn’t lead this race, other powers such as China or Russia will — and they’ll put up fewer obstacles and be less cautious about those risks. Along these lines, Holman Jenkins, a columnist for The Wall Street Journal, wrote a few days ago that Anthropic isn’t the end of the world. “Its IPO prospectus says more about national security than about human extinction,” he noted.Tough competition amid growthOne of Anthropic’s challenges is maintaining its growth rate in a scenario where AI can become cheaper as competition intensifies. Open‑weight models are narrowing the gap with leading closed models, increasing pressure on prices. The company itself has acknowledged facing tough competition, with Google and OpenAI as primary rivals.The firm said the market for AI models, platforms and applications is evolving rapidly, is highly competitive and is subject to frequent and significant changes in technology, customer needs and expectations, industry standards, and regulation. It also warns of new entrants and the emergence of new models, platforms, and applications. “Large companies that are currently our customers may choose to develop their own AI capabilities internally rather than purchase our models,” Anthropic says.Matías García Calvo, partner in charge of data and AI at NTT Data, stresses that Anthropic is moving into applications, platforms, and products. According to the expert, the company must consider how the differentiation it currently has with AI models can translate into other areas. García Calvo refers to workflows or business processes where AI delivers real enterprise value. The NTT Data expert points to the differentiation that will come from moving AI from a question‑answer assistant to an execution engine for company processes, offering greater value by optimizing those processes and generating new business models.Iñaki Arrola, of Kfund, says the unanswered question is whether Anthropic can maintain its position over the medium-term against OpenAI’s ChatGPT, Google’s Gemini, and SpaceX’s Grok when models and products become more similar.More broadly, Anthropic has a corporate strategy to drive acquisitions that add talent and technology and to enter new areas. In April 2026 Anthropic completed the acquisition of Coefficient Bio, a computational biology company, to accelerate its life‑sciences research in a deal valued at $400 million. In May it bought Stainless Software for $295 million to improve developer experience and agent connectivity.Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition

Original Source

Read the full article at English →

KhanList aggregates and links to publicly available news content. We do not host full articles from third-party sources. Always verify important information with original sources.