Elon Musk admits his OpenAI bet backfired

Elon Musk now believes one of his earliest moves to stop any single company from dominating artificial intelligence may have backfired badly.

Elon Musk now believes one of his earliest moves to stop any single company from dominating artificial intelligence may have backfired badly. In a conversation with Zanny Minton Beddoes, editor-in-chief of The Economist, the xAI founder said he originally helped create OpenAI as a counterweight to Google, which had “more or less a monopoly on AI” at the time. Musk admitted things did not play out the way he had hoped.

He pointed to Anthropic’s split from OpenAI, and its rise into a leading AI company, as proof his original bet triggered consequences he never saw coming. It ultimately sped up the AI race rather than slowing it down. He said he was not really trying to accelerate AI, but that creating OpenAI, followed by Anthropic spinning out of it, ended up producing knock-on effects that did exactly that. His takeaway was blunt, every road in AI now leads to faster development, not less.

The comments arrive amid Musk’s ongoing legal battle with OpenAI CEO Sam Altman over the company’s shift from nonprofit research lab to commercial powerhouse.

Musk has framed his frustration in simple terms, he says that he funded a nonprofit meant to be an open-source AI company owned by the world, only to watch it become an $800 billion closed-source for-profit business. That is the opposite of what he intended his donation for.

Musk had sought $150 billion in damages from OpenAI, Altman, Microsoft, and other parties. A federal jury in Oakland ruled in May that his claims fell outside the statute of limitations, and Judge Yvonne Gonzalez Rogers dismissed the case on those grounds. Musk has said he plans to appeal to the Ninth Circuit.

Despite the feud with Altman, Musk struck a notably warmer tone about Anthropic. He suggested the people who left OpenAI to start it did so because they had lost faith in Altman’s leadership, and he praised Anthropic CEO Dario Amodei as someone genuinely principled and seriously concerned about AI’s future. He went so far as to say the company’s team passed what he called his “evil detector.”

Musk also floated a safety proposal: giving leading AI labs a window of a week or two to review rivals’ models before public release. His reasoning was that competitors checking each other, even amid personal animosity between CEOs, would actually make the process more honest rather than less.