Tesla has unveiled the name of its supercomputing cluster at the Gigafactory in Texas. The cluster, now officially known as “Cortex,” was highlighted by Elon Musk following a tour of the newly constructed facility. Musk mentioned on social media that Cortex would house approximately 100,000 Nvidia H100 and H200 chips.
These chips are essential for video training of Tesla’s neural networks, which are crucial for the company’s Full Self-Driving (FSD) program and its Optimus robot. The supercomputer will require around 130 megawatts of power and cooling this year. This demand will grow to over 500 megawatts within the next 18 months.
Recent photos from drone operator and Giga Texas observer Joe Tegtmeyer have shown large fans installed in the building to handle the extensive cooling requirements. Additionally, four giant water tanks have been set up on the second floor to augment the cooling efforts. Musk has stated that Tesla will spend around $10 billion on AI-related expenditures this year, half of which will be used internally.
Tesla’s Cortex supercomputer unveiled at Gigafactory
This budget will cover the in-house-designed AI inference computer, Tesla cars’ sensors, and the Dojo supercomputer’s development. Dojo will be critical in Tesla’s future, particularly autonomous driving technologies.
Musk believes that Dojo’s advanced computing power will greatly enhance Tesla’s self-driving systems’ capabilities, making them safer and more efficient. This push towards stronger AI capabilities is part of Tesla’s broader strategy to dominate the autonomous vehicle market. The introduction of the robotaxi, which is expected to operate with minimal human intervention, will mark a significant milestone in this journey.
Tesla’s AI team has been working diligently to ensure that Dojo can handle the vast amounts of data that the robotaxi and other autonomous systems will generate. This data processing capability will be crucial for the real-time decision-making necessary for safe autonomous driving. Industry analysts are closely watching Tesla’s progress with Dojo.
Success with this supercomputer could give Tesla a significant edge in the fiercely competitive automotive industry, where major companies are racing to develop reliable and efficient autonomous driving solutions.
