When it comes to controversy, Tesla CEO Elon Musk is no stranger. From smoking weed with Joe Rogan to hot takes about immigration and the border, Musk has found himself at the focal point of many different imbroglios. But lately, the focus has been on some rather touchy leadership decisions at Tesla which have brought his company loyalties into question.
Welcome to Critical Materials, your daily roundup for all things EV and automotive tech. Today, we’re chatting about Musk’s recent diversion of nearly half a billion dollars of AI training GPU shipments from Tesla to his privately-owned companies, Stellantis’ consideration of in-house EV component development, and Mercedes-Benz tapping a self-driving startup from China for its partial-autonomy needs. Let’s jump in.
30%: Shareholders Slam Tesla CEO For Diverting Nvidia GPUs To His Private Companies
Elon Musk at Tesla’s Shareholders Meeting at Texas Gigafactory
Many Tesla shareholders who will soon weigh in on Elon Musk’s $56 billion pay package are once again upset over the CEO’s actions. This time, it’s not moving the company away from a $25,000 mass-market car in favor of robotaxis—instead, it’s the diverting a shipment of AI-training Nvidia GPUs intended for Tesla to two of his own private companies.
News broke of Musk’s decision earlier this week in the form of leaked internal emails from graphics card manufacturer Nvidia, in a scoop published by CNBC. The CEO reportedly ordered Nvidia to allow X and xAI to jump the line and receive shipments of around 12,000 Nvidia H100 GPUs originally destined for Tesla as part of a $500 million purchase commitment by the automaker, effectively delaying Tesla’s potential delivery date by “months,” according to CNBC.
Here’s a quick snippet from Nividia’s memo, which CNBC says was circulated in December:
Elon prioritizing X H100 GPU cluster deployment at X versus Tesla by redirecting 12k of shipped H100 GPUs originally slated for Tesla to X instead. In exchange, original X orders of 12k H100 slated for Jan and June to be redirected to Tesla.
Musk confirmed the redirection of the chips with the justification that Tesla had nowhere to switch on the processors by the time it would receive them:
Now, this might not seem like a big deal since Tesla is an automaker. I mean, why else would Tesla need a large number of GPUs? It’s not like its CEO declared the company to be an AI company just over a month ago and said that it would need to scale from 35,000 to 85,000 of these GPUs by the end of the year. Oh, wait.
Tesla uses these GPUs to train AI data—it’s what takes all of the clips recorded while Tesla vehicles are driving on the road and teaches the learning model how to drive a car. And while Musk has said that Tesla was “no longer […] compute constrained,” the company still committed to more than doubling its stock of H100 GPUs for training data—something critical as it plans to show off a driverless robotaxi in about two months.
Many shareholders and commentators have responded in anger.
Some are doubling down on the position that Musk has become a “distracted” CEO due to external commitments with other companies, and others are escalating the claim to breach of fiduciary duties and conflict of interest. Some have even called for a lawsuit.
Now, if Musk’s claims are true and Telsa really did have nowhere to store these GPUs, it seems to make sense to decline delivery. After all, why have a warehouse full of high-demand, depreciating hardware? But if Tesla experiences can’t meet AI targets due to issues scaling to 85,000 GPUs by the end of the year due to six months of chip shipments being sent to the CEO’s privately-held companies, shareholders will almost undoubtedly look to this as part of the cause.
The last thing that Musk needs is another target painted on him during a time when shareholder support is so critical. Investor voting will soon end for Musk’s $56 billion compensation plan, leaving the fate of his pay package in their hands—and if Musk’s leadership decisions continue to get called into question, it may not bode well for the future of autonomy and robotics at Tesla.
60%: Stellantis Could Move EV Component Production In-House Over Slow Suppliers
Carlos Tavares, Stellantis CEO
While the U.S. has taken the protectionist approach against cheap, tech-rich EVs coming out of China, the same can’t be wholly said for Europe. Vehicles built by some China-based automakers will soon begin to hit the roads in the EU, which is striking fear in just about every automaker that does business in those markets.
The solution? Rapid development and cost-cutting, done more in-house. The only problem is that suppliers are dragging their feet, at least according to Stellantis CEO Carlos Tavares.
“When suppliers are not racing at the same speed of our teams, our teams see a big benefit to in-source,” said Tavares during an investor conference call, according to Automotive News. “You come to the conclusion that what you have outsourced, you can do it in-house.”
Currently, Stellantis procures parts from big-name suppliers like Apriv, Continental, Forvia, Magna International, Melfi, and many more.
Now, Tavares hasn’t outright said which parts that it could in-house production of, nor if the automaker has committed to that approach. For all we know, it could be a public scare tactic used to light a fire under suppliers who might be taking too long to deliver components needed for Stellantis to quickly scale its EV business while Chinese rivals are encroaching on the brand’s territory with cheaper vehicles.
There was a rather coinciding announcement, however, that could point the finger toward batteries as being the source of Tavares’ concern.
Automotive Cell Company, the joint venture between Stellantis and Mercedes-Benz, announced that it would pause construction of two multi-billion-dollar battery plants in Germany and Italy. The group said that it needed time to research and develop low-cost batteries for EVs and would look to resume construction in late 2024 or early 2025. Its third plant in France is already open and not affected.
“We are going to adjust our investment plans on EVs to the pace at which market sales of EVs grow,” said Tavares. “We do not control that speed.”
90%: Mercedes-Benz Taps Chinese Startup For Future Autonomy Integrations
Mercedes CLA Concept 2023
Mercedes-Benz will tap a Chinese autonomy startup called Momenta for its next-gen driver assistance systems, according to local news outlet 36kr.
The system will reportedly debut in the 2025 Mercedes-Benz CLA EV which is said to be packed full of existing EQS tech but will use Momenta’s autonomy solution behind the scenes.
According to those familiar with the vehicle, the CLA’s autonomy stack will include Lidar—which was already confirmed by Mercedes—alongside cameras, millimeter-wave radar, and Nvidia’s Orin system-on-chip (already used by automakers like Lucid and BYD) to achieve an “enhanced” version of Level 2 partial-autonomy.
Mercedes already has an existing hands-off, eyes-off Level 3 solution, coined Drive Pilot, which is available on certain S-Class and EQS vehicles operated under certain conditions in California and Nevada. It’s not clear at this time why the automaker is forking from the existing Drive Pilot system, however, Momenta’s stack will reportedly enable the CLA to have navigation-assisted driving in dense urban environments without the need for high-definition mapping.
Now, Mercedes is no stranger to Momenta. The automaker has contributed to the startup’s corporate fundraising efforts since at least 2021. In 2017, the brand participated in Series B funding while still under the Daimler marque. This will, however, be the first time that the brand taps a company based in China for its partial-autonomy needs.
The new Mercedes-Benz CLA EV will reportedly go into production with Momenta’s autonomy stack beginning in April 2025.
100%: What Could Happen If Musk’s Pay Package Fails The Shareholder Vote?
Elon Musk at CodeCon 2021
Musk’s compensation package is a hot topic right now. Despite a huge push from Tesla’s board encouraging a positive vote, it may not turn out that way. And large shareholder groups are making it clear which way they are voting—enough to get the public attention of the CEO.
Musk has expressed his desire to build AI and robotics systems outside of Tesla if he can’t get 25% voting power of Tesla. While he has a good excuse for the diversion of Nvidia GPUs, one could argue that it feels very much like a contingency plan should his pay package fail to pass.
It’s hard to picture what Tesla would look like without Musk behind the reins, though. And it seems unlikely that Musk will leave Tesla should the package fail to pass. However, some shareholders have expressed concerns that Musk is distracted and isn’t even pretending to act like Tesla is his top priority.
All of that being said—the vote comes to a head next week. Let me know in the comments what you think will happen, and what the consequences for Tesla (and Musk) could be.
Tesla shareholders were left astounded and perplexed by CEO Elon Musk’s recent directive to divert shipments of Nvidia chips from the electric vehicle manufacturer to a yet-to-be-revealed company, known only as X.
The move, which was announced in an internal memo obtained by Bloomberg, has stirred up speculation and concern among Tesla investors. Many are questioning the rationale behind Musk’s decision to divert crucial components that are essential for the production of Tesla’s vehicles.
Nvidia chips are a key component in Tesla’s automotive software systems, contributing to the company’s autonomous driving capabilities and overall performance. With the global shortage of semiconductors already causing production delays for many automakers, diverting Nvidia shipments could further exacerbate supply chain challenges for Tesla.
Musk’s decision to redirect these critical components to an unknown entity only adds to the confusion and uncertainty surrounding the situation. The secrecy surrounding company X has led to widespread speculation about its identity and the nature of its relationship with Tesla.
Some analysts have suggested that Musk’s directive could potentially be tied to his ambitious goals for Tesla’s self-driving technology. By diverting Nvidia chips to another company, Musk may be seeking to accelerate the development and implementation of advanced autonomous driving features in Tesla vehicles.
However, the lack of transparency about the decision has left shareholders feeling unsettled and concerned about the potential impact on Tesla’s production and financial performance. Without a clear explanation from Musk or Tesla management, investors are left to speculate about the implications of this unusual move.
As Tesla continues to navigate challenges in the semiconductor supply chain, the diversion of Nvidia chips raises further questions about the company’s strategic priorities and the potential risks facing its business. Shareholders will be closely monitoring developments in the coming weeks to gain a better understanding of the situation and its implications for Tesla’s future.