Redazione RHC : 18 August 2025 10:26
On July 23, 2025, Tesla held its second-quarter earnings conference call. Elon Musk, as usual, conveyed his infectious optimism to Wall Street. Speaking about Dojo, the company’s carefully constructed artificial intelligence supercomputer, he expressed confidence: “We expect Dojo 2 to be operational at scale next year, with a capacity equivalent to approximately 100,000 H100 chips.”
This statement was a real confidence booster. Investors considered Dojo not only the technological cornerstone of Tesla’s fully autonomous driving system (FSD), but also the main driver of its transformation from an electric car company to a trillion-dollar artificial intelligence giant. Analysts at Morgan Stanley even estimated its potential value at $500 billion.
However, no one could have predicted that this “spectacular” dream would end so quickly. Just three weeks later, the storm arrived. In early August 2025, a Bloomberg report rocked the tech world: the Dojo project team had been disbanded, and its leader, Peter Bannon, was about to leave. Then, over the weekend of August 9–11, Musk personally brought this once-stellar project to a close with a series of posts on his social media platform, X. He described this dramatic turn of events as a thoughtful strategic update, declaring that Dojo 2 had reached an “evolutionary dead end” and that Tesla’s future would focus on another “converged architecture” chip, called AI6.
Less than 20 days passed between the bold promise of “100,000 H100s” and the stark conclusion of an “evolutionary dead end.”
What happened behind the scenes?
This wasn’t just a project adaptation; it was a Silicon Valley drama, intertwined with technological gambles, talent frenzy, commercial realities, and maneuverings between giants.
The failure of Project Dojo is not only a major setback for Tesla’s AI strategy, but also a multibillion-dollar warning to all ambitious individuals seeking to challenge the supremacy of hardware in the AI era.
Redazione