Archive — US — techcrunch
Tesla’s Dojo, a timeline
Elon Musk has shifted Tesla's focus from its Dojo supercomputer for AI training to the new Cortex supercluster, a move intended to enhance the company's autonomous driving technology. Despite years of teasing and development, Dojo has been officially shut down as Tesla consolidates resources towards the next-generation AI6 chips to streamline AI capabilities by 2025.
Key points
- Dojo was initially announced in 2019 as a supercomputer to improve Tesla's FSD technology.
- Musk touted Dojo's potential to process large amounts of training data to enhance automated driving.
- Years of development and teases led to a partial rollout in 2023, but challenges in Nvidia hardware availability drove Tesla to rethink the project.
- By mid-2024, Musk acknowledged a shift in focus towards the Cortex supercomputer, which utilizes a different architecture.
- In August 2025, Tesla disbanded the Dojo team, confirming the project's end and a pivot to the AI6 chips, indicating an evolutionary dead end for Dojo.
Why it matters
- Tesla's trajectory mirrors the broader trend of AI integration in automotive technologies seen across the industry.
- Recent advancements in AI architecture align with the 2025 IT trend of focusing on integrated and scalable AI solutions.
- The competition in AI hardware, especially with Nvidia’s dominance, reflects the challenges that companies face in tech innovation.
- Musk's strategy highlights the increasing corporate focus on developing proprietary technology rather than relying on external suppliers.
The shutdown of Tesla's Dojo reflects a strategic pivot towards more integrated AI solutions, emphasizing the importance of technological evolution in the rapidly advancing field of autonomous driving.