News
Tesla investing over $500 million to build Dojo supercomputer in New York
Tesla is set to invest over $500 million into building a Dojo supercomputer at its New York Gigafactory, as announced during a state budget briefing this week.
New York Governor Kathy Hochul held a hearing on economic developments in the state on Friday, officially announcing that Tesla plans to build its next Dojo supercomputing cluster at its Gigafactory in Buffalo (via WGRZ). Hochul said the project would account for an investment of $500 million, and it will also come alongside a separate artificial intelligence (AI) supercomputer being built at the State University of New York (SUNY).
Tesla DOJO predicted to bring the company’s enterprise value up to $500 Billion: report
“I’m also proud to announce that Tesla is investing $500 million to build their next supercomputer right here in Buffalo,” Hochul said.
Tesla CEO Elon Musk later went on to confirm the plans to build the supercomputing cluster on X, though he also clarified that the $500 million amount would only account for a smaller Nvidia system, adding that the “table stakes” for being a competitive player in AI were “at least several billion dollars per year.”
The governor is correct that this is a Dojo Supercomputer, but $500M, while obviously a large sum of money, is only equivalent to a 10k H100 system from Nvidia.
Tesla will spend more than that on Nvidia hardware this year. The table stakes for being competitive in AI are at…
— Elon Musk (@elonmusk) January 26, 2024
Currently, Tesla’s New York Gigafactory builds the company’s solar panels, solar roof, and the electrical components for its Superchargers, while the site also has a number of staff dedicated to Autopilot data analysis.
She didn’t provide a timeline for the installation plans, though the New York Governor also detailed $275 million in funding from a state grant as well as over $400 million in public and private funding all going toward the “Empire AI” program at SUNY.
In addition, Hochul says the global AI market has reached a total value of about $150 billion, though it’s expected to balloon to $1.3 trillion in just a few years.
The news comes after Tesla Dojo project lead Ganesh Venkataramanan left the company last month, and after Musk said last June that Dojo had already “been online and running useful tasks for a few months.”
The Dojo supercomputer is expected to be able to process massive volumes of sensor data to help train AI on real-world driving footage. Beyond Tesla’s Full Self-Driving (FSD) beta systems, Dojo is expected to have several potential applications, being one of the world’s most prolific computing clusters.
You can watch the full video of the briefing below from WGRZ, with Hochul’s statements on Tesla around 33 minutes minutes.
What are your thoughts? Let me know at zach@teslarati.com, find me on X at @zacharyvisconti, or send your tips to us at tips@teslarati.com.
News
Tesla Cybercab display highlights interior wizardry in the small two-seater
Photos and videos of the production Cybercab were shared in posts on social media platform X.
The Tesla Cybercab is currently on display at the U.S. Department of Transportation in Washington, D.C., and observations of the production vehicle are highlighting some of its notable design details.
Photos and videos of the production Cybercab were shared in posts on social media platform X.
Observers of the Cybercab display unit noted that the two-seat Robotaxi provides unusually generous legroom for a vehicle of its size. Based on the vehicle’s video, the compact two-seater appears to offer more legroom than Tesla’s larger vehicles such as the Model Y, Model X, and Cybertruck.
The Cybercab’s layout allows Tesla to dedicate nearly the entire cabin to passengers. The vehicle is designed without a steering wheel or pedals, which helps maximize interior space.
Footage from the display also highlights the Cybercab’s large center screen, which is positioned prominently in front of the passenger bench. The display appears intended to provide entertainment and ride information while the vehicle operates autonomously.
Images of the vehicle also show an additional camera integrated into the Cybercab’s C-pillar. The extra camera appears to expand the vehicle’s field of view, which would be useful as Tesla works toward fully unsupervised Full Self-Driving.
Tesla engineers have previously explained that the Cybercab was designed to be highly efficient both in manufacturing and in operation. Cybercab Lead Engineer Eric E. stated in 2024 that the Robotaxi would be built with roughly half the number of parts used in a Model 3 sedan.
“Two seats unlocks a lot of opportunity aerodynamically. It also means we cut the part count of Cybercab down by a substantial margin. We’re gonna be delivering a car that has roughly half the parts of Model 3 today,” the Tesla engineer said.
The Tesla engineer also noted that the Cybercab’s cargo area can accommodate multiple golf bags, two carry-on suitcases, and two full-size checked bags. The trunk can also fit certain bicycles and a foldable wheelchair depending on size, which is quite impressive for a small car like the Cybercab.
Elon Musk
Elon Musk’s xAI wins permit for power plant supporting AI data centers
The development was reported by CNBC, citing confirmation from the Mississippi Department of Environmental Quality (MDEQ).
Mississippi regulators have approved a permit allowing Elon Musk’s artificial intelligence company xAI to construct a natural gas power plant in Southaven. The facility is expected to support the company’s expanding AI infrastructure tied to its Colossus data center operations near Memphis.
The development was reported by CNBC, citing confirmation from the Mississippi Department of Environmental Quality (MDEQ).
According to the report, regulators “voted to approve the permit” of xAI subsidiary MZX Tech LLC to construct a power plant featuring 41 natural gas-burning turbines “after careful consideration of all public comments and community concerns.”
The Mississippi Department of Environmental Quality stated that the permit followed a regulatory review process that included public comments and community input. Jaricus Whitlock, air division chief for the MDEQ, stated that the project met all applicable environmental standards.
“The proposed PSD permit in front of the board today not only meets all state and federal permitting regulations, but goes above and beyond what is required by law. MDEQ and the EPA agree that not a single person around our facilities will be exposed to unhealthy levels of air pollution,” Whitlock stated.
The planned facility will help provide electricity for xAI’s AI computing infrastructure in the Memphis region.
The Southaven project forms part of xAI’s efforts to scale computing capacity for its artificial intelligence systems.
The company currently operates two major data centers in Memphis, known as Colossus 1 and Colossus 2, which provide computing power for xAI’s Grok AI models. xAI is also planning to build another large data center in Southaven called Macrohardrr, which would be located in a warehouse previously used by GXO Logistics.
Large-scale AI training requires substantial computing power and electricity, prompting technology companies to develop dedicated energy infrastructure for their data centers.
SpaceX President Gwynne Shotwell previously stated that xAI plans to develop 1.2 gigawatts of power capacity for its Memphis-area AI supercomputer site as part of the federal government’s Ratepayer Protection Pledge. The commitment was announced during an event with United States President Donald Trump.
“As part of today’s commitment, we will take extensive additional steps to continue to reduce the costs of electricity for our neighbors. xAI will therefore commit to develop 1.2 GW of power as our supercomputer’s primary power source. That will be for every additional data center as well. We will expand what is already the largest global Megapack power installation in the world,” Shotwell said.
“The installation will provide enough backup power to power the city of Memphis, and more than sufficient energy to power the town of Southaven, Mississippi where the data center resides. We will build new substations and invest in electrical infrastructure to provide stability to the area’s grid.”
Elon Musk
Tesla China teases Optimus robot’s human-looking next-gen hands
The image was shared by Tesla AI’s account on Weibo and later reposted by Tesla community members on X.
A new teaser shared by Tesla’s China team appears to show a pair of unusually human-like hands for Optimus.
The image was shared by Tesla AI’s account on Weibo and later reposted by Tesla community members on X.
As could be seen in the teaser image, the new version of Optimus’ hands features proportions and finger structures that look strikingly similar to those of a human hand. Their appearance suggests that they might have dexterity approaching that of a human hand.
If the image reflects a new generation of Optimus’ hands, it could indicate Tesla is continuing to refine one of the most critical components of its humanoid robot.
Hands are widely viewed as one of the most difficult engineering challenges in robotics. For Optimus to perform complex real-world work, from manufacturing tasks to household activities, its hands would need to be the best in the industry.
Elon Musk has repeatedly described Optimus as Tesla’s most important long-term product. In posts on social media platform X, Musk has stated that Optimus could eventually become the first real-world Von Neumann machine.
In theory, a Von Neumann machine is a self-replicating system capable of building copies of itself using available materials. The concept was originally proposed by mathematician John von Neumann in the mid-20th century.
“Optimus will be the first Von Neumann machine, capable of building civilization by itself on any viable planet,” Musk wrote in a post on X.
If Optimus is expected to carry out complex work autonomously in the future, high levels of dexterity will likely be essential. This makes the development of advanced robotic hands a key step towards Musk’s long-term expectations for the product.