Elon Musk has expressed a bold vision for Tesla, aiming to establish the company as a leader in artificial intelligence (AI) and robotics. To realize this goal, Musk emphasized the need for substantial investments in Nvidia’s high-performance processors, crucial for enhancing Tesla’s infrastructure. During Tesla’s first-quarter earnings call in April, Musk revealed plans to significantly increase Tesla’s use of Nvidia’s H100 chips, from 35,000 to 85,000 by the end of the year. He also announced that Tesla would allocate $10 billion toward AI training and inference over the same period.
However, emails from senior Nvidia staff, which have been widely circulated within the company, suggest that Musk may have overstated Tesla’s procurement plans. The correspondence reveals that a significant portion of the Nvidia chips originally intended for Tesla was diverted to Musk’s social media company, X, formerly Twitter. Musk reportedly prioritized a shipment of AI processors meant for Tesla, redirecting more than $500 million worth of Nvidia’s graphics processing units (GPUs) to X, which caused delays in the automaker’s plans to develop autonomous vehicles and humanoid robots. An internal Nvidia memo from December stated that 12,000 of the H100 chips originally designated for Tesla were rerouted to X, which postponed Tesla’s expected receipt of the chips for months.
Further emails, from late April, highlighted discrepancies between Musk’s statements and Nvidia’s records, suggesting that his claims about Tesla’s AI spending and chip bookings did not align with actual data. These emails also warned that ongoing layoffs at Tesla could hinder progress on projects like the H100 supercomputing initiative at Tesla’s Texas Gigafactory, further delaying efforts to build the necessary infrastructure for AI and robotics development.
This new information underscores growing tensions between Musk and some of Tesla’s investors, who are concerned about Musk’s ability to balance the demands of multiple companies. Besides leading Tesla, Musk is also the CEO of SpaceX, the founder of Neuralink, and the head of The Boring Co. He also owns X and launched the AI-focused startup xAI in 2023. These ventures are all vying for limited resources, including Nvidia’s coveted AI chips, which are in high demand across the tech industry.
Tesla, struggling with declining sales, faces increased competition and pressure to deliver new products. In a recent poll, Tesla’s reputation suffered in the U.S., partly due to Musk’s controversial political remarks. Tesla’s stock price has dropped 29% this year, prompting some shareholders to question Musk’s focus on the company’s core business. Instead of addressing concerns about Tesla’s current performance, Musk has been emphasizing the company’s future products, including self-driving cars, robotaxis, and a driverless transportation network. He continues to claim that Tesla will achieve full autonomy, a promise he made again during the earnings call in April.
To make these future ambitions a reality, Musk has stressed that Tesla requires vast quantities of Nvidia’s GPUs, which are in short supply due to the growing demand from major players like Google, Amazon, Microsoft, Meta, and OpenAI. Nvidia CEO Jensen Huang has acknowledged the challenge of meeting this demand, with the company’s GPUs being “consumed every which way” by data centers and AI development projects.
Despite the scarcity of these chips, Musk has promoted his companies’ plans for large-scale AI infrastructure. At Tesla, he has promised to build a $500 million supercomputer in Buffalo and a high-performance computing cluster in Texas. These systems are expected to power Tesla’s efforts in developing computer vision and large language models (LLMs) necessary for robotics and autonomous vehicles. Musk has also committed to building “the world’s largest GPU cluster” at xAI’s North Dakota facility, further exacerbating the competition for Nvidia’s processors.
In January, Musk acknowledged that Tesla’s ambitions in AI and robotics would require more control over the company, hinting that he might seek increased voting rights. Despite owning 20.5% of Tesla’s shares, Musk has expressed discomfort with his current level of influence and suggested that he would prefer to build products outside of Tesla if he cannot gain more control. This statement has raised concerns among long-time Tesla investors, including Leo Koguan, Tesla’s largest retail shareholder, who criticized Musk’s demands as “blackmail.”
Experts have pointed out that Musk’s decision to prioritize his private companies over Tesla when it comes to acquiring critical hardware raises potential conflicts of interest. Joel Fleming, a securities litigator, explained that when a CEO oversees multiple companies competing for the same resources, it can lead to corporate opportunities being diverted away from one company in favor of another. Musk’s handling of this situation, including his decision to use Tesla’s chips for X, has drawn criticism for potentially undermining Tesla’s progress on AI and robotics.
Musk has a history of blending resources across his companies. For example, after acquiring Twitter, Musk utilized Tesla engineers to help overhaul the platform, with some employees even working for both Tesla and X simultaneously. At xAI, Musk has also recruited former Tesla employees, including those involved in Tesla’s Autopilot and data projects.
This intertwining of resources, especially the redirection of Nvidia chips away from Tesla, has raised serious concerns among investors and experts alike. While Musk continues to promote his grand vision for AI and robotics, the delays in Tesla’s supercomputing projects and the growing competition for essential hardware could ultimately hinder the company’s progress toward these ambitious goals.