Latest

Artificial Intelligence

IBM caught between Anthropic, the AI it helped train, and the Pentagon

IBM is left trapped between Anthropic and the Pentagon after training the AI that threatens its COBOL business and raises doubts on Wall Street

IBM caught between Anthropic, the AI it helped train, and the Pentagon
IBM caught between Anthropic, the AI it helped train, and the Pentagon

Wall Street's financial community is talking about nothing else. How could a giant with a century of experience like IBM make such a costly miscalculation by joining forces with Anthropic. Last Monday's 13.15% stock plunge is, beyond a financial punishment, the result of an alliance that, in barely 140 days, has proved disastrous.

IBM opens its doors to Anthropic

On October 7, 2025, IBM opened the doors of its most exclusive and hermetic clients, the world's major banks, to Anthropic's engineers. The goal was for Anthropic's AI (Claude) to learn the complexities of critical systems. However, Anthropic used that privileged learning to launch Claude Code, a tool that automates the modernization of COBOL, the language IBM has squeezed for decades through slow and expensive consulting services.

Faced with the embarrassment of having been deceived by its own partner, Rob Thomas, IBM's head of Software, came out in defense of the integrity of its systems. In an attempt to calm investors, Thomas explained that IBM's value is not the code, but the box it runs in. "Translating code is one thing. Modernizing a platform is something entirely different. The value of the mainframe has nothing to do with COBOL, it has to do with the platform, an architecture designed from the silicon up to the operating system for unmatched security and efficiency. The language is not the source of that value."

The Pentagon

Despite this institutional defense, the market delivered its verdict. If Anthropic's AI can do the dirty work of COBOL, the need to pay IBM's consulting tax is drastically reduced. To make matters worse, while IBM tries to defend its business against Anthropic, the Pentagon has put the startup on its target list. Defense Secretary Pete Hegseth has given Anthropic an ultimatum that expires this Friday, telling it to either remove its ethical red lines (filters against autonomous weapons and mass surveillance) or the Government will invoke the Defense Production Act to intervene in the startup.

If the Pentagon carries out its threat to label Anthropic a security risk for not giving up its ethical filters, any company using Anthropic technology will be barred from working with the U.S. Government. IBM has billion-dollar contracts with the Department of Defense and other federal agencies. If its watsonx platform contains Anthropic technology, IBM would be forced to choose. Either it breaks with Anthropic and loses its best AI, or it loses all its Government contracts.

Nicolás Maduro

Should the Government invoke this law, it could force Anthropic to prioritize military orders over civilian ones. This would affect IBM, because the tech company has sold its clients, such as banks and insurers, on the idea that its AI is private and secure. If the Pentagon takes control, IBM's clients will fear that the military has a backdoor to spy on their data. No one will want to use an AI controlled by generals.

Tension exploded after the operation against Nicolás Maduro on January 3, when it leaked that the military used Claude (Anthropic's AI) to coordinate live-fire operations. Anthropic was outraged and demanded explanations, which infuriated the Pentagon. That event was the point of no return that broke the trust between Silicon Valley and the Pentagon.

The Government wants AI companies to stop asking questions about how their software is used. For IBM, this is a reputational disaster, since its model is based on governance and transparency, something the Pentagon wants to eliminate entirely. It should be noted that Anthropic is not a conventional company. It is what is called a Public Benefit Corporation (PBC), which means it can legally put ethics above profits.

Perfect storm

Its owners are divided into three groups. On one side are the founders, the Amodei siblings, Dario, the CEO, and Daniela, the president. Both founded the company in 2021 after leaving OpenAI, because they felt Sam Altman was neglecting safety. Along with five other co-founders, they own approximately 10% or 12% of the company, which has made them billionaires after the recent funding round closed on February 12.

In that operation, led by funds such as GIC and Coatue, Anthropic raised 30 billion dollars, reaching an official valuation of 380 billion dollars. Amazon and Google remain the largest outside investors. Amazon has invested about 8 billion dollars and Google another 2 billion. Although they do not control day-to-day operations, they do provide the cloud infrastructure where the AI lives.

The summary is a perfect storm for Big Blue: its technology partner snatches away the COBOL business while its main political ally, the Pentagon, could disqualify it. IBM is left alone at a crossroads where the only applause it gets is moral support from figures like Ethereum creator Vitalik Buterin. Scant consolation for a company watching its century-old model disintegrate before an AI it itself helped train.

Covadonga Fernandez

Author

Covadonga Fernandez

Covadonga Fernández is founder and director of the newspaper Observatorio Blockchain. She has worked with various companies developing crypto/blockchain projects, such as Singular DTV, a Switzerland-based platform and a pioneer in decentralizing the film industry, which led her to take part in the 2019 San Sebastián Film Festival. That same year she also took part in the Bern International Film Innovation Festival (Switzerland), with a presentation on linguistic plurality in blockchain. In 2022 she took part in the Acción Cultural Española edition with an article on Metaverse and Culture. In 2017 she wrote the chapter on Blockchain and Journalism in the book: «Blockchain: The Industrial Revolution of the Internet», published by Grupo Planeta. A year earlier, in 2016, she created the first Meetup on Blockchain and Journalism. That same year she also created the Blockchain Media site, to track the decentralization process of the cultural and creative industries. In 2020 she co-organized an event on how blockchain could help the cultural and creative industries, where the money received from sponsors was tokenized and speakers received payment in NFT format. Before founding Blockchain Media and Observatorio Blockchain, Covadonga worked at Grupo Z and the newspaper ABC; she was communications director of the Madrid Chamber of Commerce, a journalism professor at Universidad Carlos III de Madrid and chair of Telemadrid. Winner in the Outreach category at the 2024 Financial Innovation Awards and honored by Alastria at the 2025 Blockchain Awards

The latest
From Huang to Botín, from Zara to Byron: corporate responsibility in global capitalism

Keep reading