It has all the ingredients of a Hollywood script: two ambitious tech nerds locked in a high-stakes battle for dominance in artificial intelligence (AI), complete with surprising plot twists.
Take Dario Amodei, CEO of Anthropic, who recently called on the AI industry to pause development, warning that humanity could lose control of the technology.
What made the statement remarkable was its timing: just days earlier, his company had filed paperwork with the US securities regulator, the SEC, to launch an IPO — a first public sale of company shares on the stock market. Anthropic’s move caught rival OpenAI off guard. ChatGPT’s owner only announced its own plans to go public a week later.
The timing looks well-chosen. Stock markets are buoyant, and AI is everywhere. Anthro at $852 billion (€738 billion). A successful IPO could push both into the exclusive club of trillion-dollar companies, a league currently occupied by only a handful of firms, including Nvidia, Apple, Alphabet, Amazon, Meta and Tesla. For context: Siemens, Germany’s largest publicly listed company, is valued at around $230 billion.
Where does all the money go?
Research firm Gartner estimates that global AI spending will exceed $2.5 trillion this year alone, with the biggest share going toward infrastructure, chiefly the construction and leasing of large data centers that provide the raw computing power AI systems require.
Until now, both companies have funded themselves through private investment rounds, attracting capital from corporations and venture funds that bet on their future success. According to Harrison Rolfes, an analyst at PitchBook, a US-based provider of data and research, OpenAI has raised $186 billion since its founding; Anthro
Anthro
Most financial analysts give Anthrotory right now, and the numbers make the case,” Rolfes told DW. Anthroll ahead of OpenAI’s projected $30 billion, something it will achieve with less overall capital raised.
Anthro1,000 enterprise customers are each spending more than $1 million annually” on Anthropic, Rolfes noted.
OpenAI, by contrast, dominates the consumer market through ChatGPT, which boasts over 900 million weekly users. The problem: most of them use it for free. “Monetizing a free user base at scale is a fundamentally different and more difficult problem,” said Rolfes.
Pedro Domingos, emeritus professor of computer science at the University of Washington, largely agrees. “They’re in the lead with business customers, where the most money is likely to come from. But things are still very fluid.”
Anthro “Maybe OpenAI should sell compute to Anthropic, but [the two companies] hate each other too much for that,” Domingos told DW.
For now, both firms rely heavily on computing resources borrowed from larger partners. OpenAI leans on Microsoft, which holds a stake in the company. Anthroossus data center.
Amodei and Altman: Two egos, one race
The rivalry is also deeply personal. In 2021, Dario Amodei left OpenAI because he was unhappy with the direction Sam Altman was taking the company in — too much focus on profit, not enough on responsibility. Since founding Anthropic, Amodei has positioned himself as a cautious voice in the industry, pushing for strong AI regulation to guard against risks.
He also drew a firm line with the US military: Claude, Anthropic’s AI, would not be used for mass surveillance of American citizens or autonomous weapons systems. The Pentagon responded by classifying Anthrosupply chain security risk,” a designation normally reserved for foreign companies.
Sam Altman has moved to fill the gap. OpenAI software is now set to be deployed by the Pentagon, and Altman’s company is increasingly cast as the industry’s “bad guy.”
That’s a striking reversal. When OpenAI was founded in 2015 as a non-profit, responsible AI development was its defining mission — a vision compelling enough to attract top engineers away from Google.
<figure class="placeholder-
Some experts suspect that Amodei’s cautious public stance is, at least partly, a marketing strategy. Domingos believes the pressures of rapid growth will eventually test Anthropic’s self-
“They’ll have to make hard choices and adapt,” for example, on the use of AI for military purposes, he said. Anthropic’s CEO “will ultimately do what needs to be done, and some Anthro Amodei and others once left OpenAI.
The race for AGI
At its core, said Domingos, both companies are chasing the same goal: Artificial General Intelligence, or AGI — a system capable of performing any intellectual task as well as or better than a human can.
Both companies’ CEOs “think it’s close, and whoever gets there first will have such an advantage that they’ll sweep the board.” Domingos called this reasoning “dubious, but it’s what they believe.”
PitchBook’s Rolfes urges caution about declaring any winner too soon.
“Getting there ‘first’ doesn’t win the battle,” he said. Turning AI into lasting profit requires broad adoption, the trust of enterprise customers, and solid margins. “The real battleground isn’t ChatGPT versus Claude, it’s which AI engine gets embedded inside the world’s largest companies.”
This article was translated from German.
