Tesla Shuts Down Dojo AI Supercomputer
Tesla Shifts AI Strategy, Scales Back Dojo in Favor of Nvidia and Samsung
date”>October 27-29, 2025
Tesla is recalibrating its ambitious AI plans, signaling a significant shift away from its in-house Dojo supercomputer project and a greater reliance on established tech giants like Nvidia and Samsung. This strategic pivot comes after years of progress and promises surrounding Dojo, initially envisioned as a cornerstone of Tesla’s full self-driving (FSD) and robotics initiatives.
From Dojo Dreams to Nvidia Dependence
For years, Tesla touted Dojo as a revolutionary approach to AI training, designed specifically for the massive datasets generated by its fleet of vehicles. The project was a dual effort - building a powerful supercomputer and designing custom chips (D1 and the planned D2) to power it. At Tesla’s first AI Day in 2021, Venkataramanan presented the D1 chip, highlighting its potential alongside Nvidia GPUs. The D2 chip was intended to address bottlenecks in the D1’s architecture.Though, talk of Dojo began to fade around August 2024, coinciding with Elon Musk’s increasing enthusiasm for “Cortex,” Tesla’s “giant new AI training supercluster being built at tesla HQ in Austin to solve real-world AI.” Now, sources indicate Tesla is increasingly leaning into partnerships with Nvidia, AMD, and Samsung.
This change in direction isn’t necessarily a sign of failure, but rather a pragmatic adaptation to the rapidly evolving AI landscape. Building and maintaining a cutting-edge supercomputer and custom chip fabrication is incredibly complex and expensive. Leveraging the expertise and existing infrastructure of industry leaders allows Tesla to accelerate its AI development without bearing the full burden of these challenges.
A $16.5 Billion Bet on Samsung for AI Chips
The moast concrete evidence of this shift is Tesla’s recent $16.5 billion deal with Samsung. This agreement will see Samsung manufacture Tesla’s AI6 inference chips. These chips are designed to be versatile, powering everything from FSD and the Optimus humanoid robot to high-performance AI training in data centers.
This move represents a significant outsourcing of chip manufacturing, a departure from Tesla’s earlier ambitions of in-house chip production. It also highlights the growing importance of inference chips – the processors that use trained AI models to make predictions and decisions – in Tesla’s overall AI strategy.
Convergence and Potential Redundancies
During Tesla’s second-quarter earnings call, Musk hinted at streamlining efforts. “Thinking about Dojo 3 and the AI6 inference chip, it seems like intuitively, we want to try to find convergence there, where it’s basically the same chip,” Musk stated. This suggests Tesla is aiming to consolidate its chip development, perhaps reducing redundancy and focusing resources on a unified chip architecture.This consolidation coudl lead to workforce adjustments within Tesla’s AI teams,though the company has not publicly commented on potential layoffs. The move underscores the dynamic nature of the AI industry, where strategies are constantly being reevaluated and refined.
Securing Musk’s Leadership Amidst the AI Race
The strategic shift comes alongside Tesla’s board offering Elon Musk a $29 billion pay package.This substantial compensation is intended to ensure Musk remains focused on driving Tesla’s AI advancements, preventing him from being overly distracted by his other ventures, including xAI, a more pure-play AI startup.The board clearly recognizes the critical importance of AI to Tesla’s future and is incentivizing Musk to prioritize its development.
tesla is navigating a complex and competitive AI landscape. By embracing partnerships and streamlining its chip development, the company is positioning itself to remain a leader in autonomous driving and robotics. While the Dojo dream may have evolved, Tesla’s commitment to AI innovation remains unwavering.have a sensitive tip or confidential documents? We’re reporting on the inner workings of the AI industry – from the companies shaping its future to the people impacted by their decisions. Reach out to Rebecca Bellan at rebecca.bellan@techcrunch.com and Maxwell zeff at maxwell.zeff@techcrunch.com. For secure communication, you can contact us via Signal at @rebeccabellan.491 and @mzeff.88.