Elon Musk Shuts Down Tesla Dojo, Calls It ‘Evolutionary Dead End’
▼ Summary
– Tesla has disbanded its Dojo AI supercomputer team, shifting focus to AI5 and AI6 chips after determining Dojo 2 was an evolutionary dead end.
– The Dojo project, including the planned D2 chip, has been shelved in favor of third-party AI5 and AI6 chips from TSMC and Samsung, designed for Tesla’s self-driving and robotics systems.
– Elon Musk stated that consolidating resources into AI5/AI6 chips is more efficient, as they can be used for both inference and training while reducing network complexity.
– Tesla’s pivot follows Musk’s earlier promotion of Cortex, a new AI training supercluster, though its current status remains unclear.
– The strategy change occurs amid declining EV sales and brand challenges, with Musk emphasizing Tesla’s autonomy ambitions despite a rocky robotaxi launch.
Tesla has officially scrapped its ambitious Dojo supercomputer project, with Elon Musk declaring the technology an “evolutionary dead end” in a recent social media post. The decision marks a significant pivot in the company’s AI development strategy, shifting focus toward its next-generation AI5 and AI6 chips instead.
Musk revealed the abrupt shutdown on X, explaining that Tesla’s resources will now concentrate exclusively on refining its AI5 and AI6 hardware, manufactured by TSMC and Samsung. These chips are designed to handle both onboard inference, critical for Tesla’s Full Self-Driving (FSD) system, and large-scale AI training. The move effectively sidelines the previously hyped Dojo initiative, which had been positioned as a cornerstone of Tesla’s autonomous driving and robotics ambitions since 2019.
Originally, Tesla had planned to expand Dojo with a second-generation D2 chip and a dedicated supercomputing facility. However, Musk confirmed that developing two distinct AI architectures no longer made strategic sense. Instead, Tesla will optimize its AI5 and AI6 chips for dual purposes, integrating them into high-density computing boards to streamline network infrastructure. “One could call that Dojo 3,” Musk quipped, suggesting a conceptual evolution rather than a complete abandonment of supercomputing goals.
The decision arrives amid broader challenges for Tesla, including declining electric vehicle sales and controversies surrounding Musk’s political engagements. The company has been working to reassure investors about its long-term prospects in autonomy, despite a rocky rollout of its robotaxi service in Austin earlier this year. Reports of erratic driving behavior from these vehicles have raised questions about the readiness of Tesla’s self-driving technology.
Meanwhile, uncertainty lingers over the fate of Tesla’s $500 million Dojo facility in Buffalo, New York, as well as the status of Cortex, a separate AI supercluster Musk hinted at in mid-2024. Neither project has received recent updates, leaving industry observers to speculate whether Tesla’s AI roadmap has undergone further revisions behind the scenes.
As Tesla redirects its engineering efforts, the AI race intensifies among tech giants and automakers alike. Musk’s latest pivot underscores the volatile nature of cutting-edge technology development, where even the most promising projects can quickly become obsolete in the face of rapid innovation.
(Source: TechCrunch)
