Brief IA
›

Anthropic Raises $81.3 Billion on Nasdaq

Business & Startups·Tom Levy·

Anthropic Raises $81.3 Billion on Nasdaq

Anthropic Raises $81.3 Billion on Nasdaq
Key Takeaways
1Anthropic raised $81.3 billion during its Nasdaq debut on Thursday.
2METR estimates that AI has roughly doubled the pace of its R&D progress.
3Dustin Moskovitz is urging the lab and its competitors to slow down.
Why it matters : Anthropic is entering the stock market while maintaining more ambitious safety commitments that remain voluntary.

Anthropic raised $81.3 billion on Thursday for its debut on the Nasdaq, setting a record for an initial public offering. At the close, its valuation reached $1.76 trillion. This listing comes as its safety commitments are being questioned, even among its earliest investors.

A First Investor Calls for a Slowdown

Dustin Moskovitz, co-founder of Facebook and a long-time investor in Anthropic after investing in OpenAI, is urging the lab to slow down. He claims to have invested funds with his wife Cari Tuna to promote safety, not a race for speed. In his view, the advancements in AI are happening too quickly: his call is also directed at other leading labs.

The couple is not expected to personally benefit from the listing: a spokesperson stated that their shares were donated to a nonprofit organization in 2025. For individual investors in the stock market, however, this operation could provide access to potential profits from AI.

Robert Armstrong of the Financial Times estimates that the lab would need to exceed $600 billion in annual revenue to achieve the 20% return on invested capital obtained by its peers. Apple's revenue stands at $478 billion. Dario Amodei, CEO of Anthropic, aims to secure an investment-grade credit rating and plans new debt issuances. These funds, like the capital raised, are intended to acquire more computing capacity.

Non-Binding Safety Commitments

The AI auditor METR estimates that the use of AI for research has roughly doubled the pace of R&D advancements at Anthropic. Such a doubling would indicate that the lab has crossed the risk threshold it set for automated R&D. Its commitments would then include additional audits, safety measures, and increased transparency.

However, the Responsible Scaling Policy that contains these commitments is not legally binding. Anthropic had supported SB 53, the California law that requires companies to publish safety policies and then comply with them legally. After its adoption, the lab submitted a less stringent framework under this law, with its more ambitious promises now classified as voluntary.

Futurists have envisioned a scenario of recursive improvement: a sufficiently advanced AI could design its own successors and then repeat the process at an accelerating pace. In this scenario, humans might no longer be able to understand the system or interrupt its operation. At the Nasdaq, Dario Amodei promised to adjust the training and deployment pace to the time required for safety. He also assured that Anthropic would resist both political and commercial pressures.

La lettre du soirGratuit · 19 h

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

Protest and Statement Against Anthropic

In San Francisco, outside the lab's headquarters, over 100 people protested with signs calling for an end to technological competition or warning about dangers to the planet. Jacob Coxon, a former employee of Anthropic whose viral resignation message had reignited attention around the AI safety movement, spoke at the event. Four individuals were arrested for allegedly blocking access to the offices.

The MisAnthropic Declaration includes signatories such as Gary Marcus, a skeptic of AI, as well as Guillaume Verdon, a proponent of its acceleration known by the pseudonym Beff Jezos. They accuse Anthropic of wanting to lead in the development of self-improving AI, despite its own warnings about potential loss of control and concentration of power. They also criticize it for exploiting its safety image with employees, the public, and investors while reducing its commitments and advocating for rules favorable to its interests.

The signatories thus challenge the safety claimed by Anthropic and reject the accumulation of power justified by the common good. The protesters denounce the weakness of government intervention to stop the company.

Amodei's Vision for AI Safety

Founded by former members of OpenAI, Anthropic aims for systems whose intelligence would surpass that of humans. However, its founders have repeatedly warned about the potential for massive destruction from such technologies. In his speech at the Nasdaq, Dario Amodei invoked possible scientific advancements, reminded of the dangers of misuse, and presented the benefit for humanity as the reason for the lab's creation.

The leader argues that Anthropic's economic success, combined with its safety commitments, could lead its competitors toward similar practices. An anonymous former employee described this vision by comparing superintelligence to the One Ring of Sauron, which corrupts its bearers in The Lord of the Rings. In this depiction, the first to build it would possess cyber weapons, cures for all diseases, and tireless digital intelligences at their service. The morality of the one who holds this power would then be decisive.

Anthropic has prioritized the laws of scaling models, while competitors have explored robotics and world models. Concerns about an AI bubble also focus on the funding circuits between companies and the geopolitical consequences of Jensen Huang's activities.

With its valuation, Anthropic is now the largest public benefit corporation in the world. It is also among the largest publicly traded companies.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.