Anthropic challenges the AI boom with the "most ridiculous" stock market debut of the year

BarcelonaThe stock market debut of Anthropic, the company behind Claude, is one of the most anticipated events of the year. Just months after the record broken by thepublic market entry of SpaceX, Elon Musk's aerospace and artificial intelligence brand, the company led by Dario Amodei could post even more spectacular figures: while Musk managed to list with a valuation of 1.77 trillion dollars, everything suggests that its competitor could reach two trillion. This is anticipated by the initial public offering (IPO) document, leaked by the Reuters agency. The same agency recently pointed out that the IPO could be resolved after the midterm elections in the United States, scheduled for November 3rd.

The figures advanced in the operation's prospectus are of a scale difficult to calculate. According to Reuters, Anthropic recorded revenues of 4.6 billion dollars in 2025, 12 times more than the previous year. However, it maintained operating losses of more than 8 billion; and net losses of 42 billion. As is now usual, the most problematic figure is the commitment to spending on computing infrastructure: it expects to allocate 518 billion dollars to the cloud and data centers in the coming years. The laboratory led by Amodei, it must be said, expects exponential growth for 2026: the revenue forecast based on business until July soared to 65 billion dollars, as reported by several North American media outlets.

The figures have made investment experts tremble. In fact, in a controversial market note sent last week, one of the senior analysts at the financial research firm New Constructs branded the IPO as "the most ridiculous of 2026" due to being oversized. "Anthropic's stock market debut is an unprecedented test of investor credulity," the statement asserted. Several experts consulted expressed themselves in a similar vein, but at the same time, they do not doubt that current demand will allow Anthropic to achieve its valuation goals. "It is realistic for Anthropic to exceed a valuation of two trillion given the current market," says Javier Cabrera, an analyst at XTB.

A 30 trillion race

In fact, the figures that Anthropic is presenting to the market suggest that the operation could become central to the economy. As reported by the Wall Street Journal at the end of August, the company expected to present investors with a total addressable market (TAM) for its AI of 30 trillion dollars. This is a calculation that forecasts the total revenue that artificial intelligence can obtain once it has fully penetrated users. The estimate is in line with, albeit slightly above, the 26.5 trillion put on the table by SpaceX. If this estimate is met, the potential market for artificial intelligence would be close to the gross domestic product of the United States.

Not everyone is confident that the scope promised by Anthropic is realistic. "If you go by common sense, this market is not there," reflects Xavier Brun, head of equity at Trea AM. Even so, stock market movements continue to support the investment theses of AI companies: SpaceX's stock, which went through a roller coaster during the first weeks of trading, is moving during Monday's session at around 163 dollars, a significant gain compared to the 135 with which it entered the stock market, and in line with the price of the first day of activity. With this reference, analysts insist that, despite everything, it is not impossible to meet the forecasts: "If Anthropic is able to maintain growth, a valuation of two trillion is not far-fetched," observes the founder of WeArtificial, Aleix Valls.

The strategic dangers of AI

The months leading up to the major artificial intelligence stock market debuts have been more than turbulent: the attack by OpenAI agents on the artificial intelligence platform Hugging Face –now acquired by Nvidia for 13 billion dollars– was just the beginning. On Friday, in fact, Anthropic acknowledged that its agents had autonomously sent a fake murder report to the Philadelphia police department, and had attempted to submit several visa applications to the State Department, as reported by The New York Times.

The accumulation of incidents related to AI agents from the main labs exploded at the beginning of September, when the head of alignment at Anthropic, Evan Hubinger, acknowledged on social media following the resignation of researcher Jacob Coxon from the company that "there is more than a 10% chance that AI will kill all humans in the next decade".

That alarm cooled down as the weeks passed, and the market is beginning to view the panic unleashed by the two AI giants as a strategy. "That AI could kill us is a lie. It would be assuming it has a will, and we are not there, nor do I know if we ever will be," says Valls. In the expert's opinion, Anthropic and OpenAI are taking advantage of these incidents to demand regulation that they see as "an insurance policy" against potential legal liabilities. "If I do everything according to the law, I am not to blame for anything. The law is," he explains.

Xavier Brun, for his part, warns of a strategy to eliminate competition, attributing to new players in the AI market, especially those with open weights and open source –which everyone can use for free–, an even higher risk. "If I, who am the best, have had the AI get away from me, it means that this is not something just anyone can do," the expert narrates. Thus, they would be looking to raise the barriers for the sector to standards that only the most established companies can meet.

In Cabrera's opinion, however, it may have backfired on them: "The calculation error was the impact the crisis had," says the analyst. The alerts reached an American public with too much insistence that is no longer among those who trust AI the most. This time, public opinion about the technology may cause "voters to pressure governments" to tighten the reins on tech companies. Regulation, therefore, could "harm the sector more than favor it," as Amodei and OpenAI CEO Sam Altman were seeking, according to experts.