In 2015, OpenAI was founded by Sam Altman and Elon Musk with a lofty mission: to ensure artificial intelligence benefits all of humanity, not profit. Altman was also deeply concerned about Google’s growing dominance. In an interview, he warned, “Over time, we’ll move closer to creating technology that surpasses human intelligence, yet people remain skeptical about how much Google would truly be willing to share.” Both Musk and Altman believed open-source AI was the best way to counter Google’s monopoly.

Before becoming bitter rivals, Sam Altman and Elon Musk once shared a close and collaborative relationship.

According to documents revealed years later, on May 25, 2015, Altman sent Musk an email stating: “I keep wondering whether it’s possible to stop humanity from developing artificial intelligence. I think the answer is almost certainly no. If AGI is destined to happen, then perhaps the first organization to make it real should be one that isn’t Google.”

Altman proposed that Y Combinator could launch a “Manhattan Project” for artificial intelligence. He suggested a structure in which, through a nonprofit model, the technology would belong to the entire world. However, if the project succeeded, those involved could still be compensated in ways similar to startup incentives. He emphasized that the organization would comply with—and actively support—regulatory oversight and safety checks.

Musk replied simply: “It might be worth discussing.”

Shortly afterward, Altman suggested that he and Musk could jointly establish a nonprofit AI research lab. The lab would aim to catch up with Google in the race toward AGI, but by taking a fundamentally different approach.

In June 2015, Altman wrote to Musk again: “Our mission is to create the first AGI and empower as many people as possible—meaning a distributed version that is likely the safest. In short, our top requirement is safety. This technology will belong to the foundation and be used to ‘benefit the entire world.’” He also proposed starting with a small team of seven to ten people, then gradually scaling up.

Musk responded: “I completely agree.”

Eventually, Musk came up with the name for the new lab: the “Open Artificial Intelligence Research Institute,” shortened to “OpenAI.”

According to Musk, they also had a so-called “founding agreement,” whose core principles were that OpenAI would be a nonprofit organization, its products would be open-source, and it would compete with Google and DeepMind in the AGI race as a critical counterbalance—while prioritizing the benefit of humanity over shareholder profit.

Years later, Musk accused Altman of violating this founding agreement. Altman firmly denied the claim, insisting that no such agreement ever existed. Musk, for his part, was unable to provide evidence showing that Altman had signed or formally agreed to any such deal.