Will Tesla invest in a new type of datacenter?

Tesla service center
Sheila Zabeu -

August 18, 2023

All indications are that electric vehicle giant Tesla is considering or already planning to enter the world of datacenters. According to the Electrek website, the company has posted a vacancy for a senior professional to become a manager “responsible for leading the design and engineering of datacenters, from concept to launch”.

The job description adds that the role will lead the end-to-end design of “the first datacenters of its kind” and will be a key member of the factory’s engineering team. However, Tesla didn’t give any details about what it meant by “the first of its kind”.

Last December, in a live Twitter event, Elon Musk called Sacramento “possibly the worst place to have a datacenter”, complaining about the high temperatures in California’s inland capital.

Elon Musk, owner of Tesla CEO who acquired Twitter in October 2022 and in July 2023 rebranded the platform as X.com, has already made some moves around datacenters associated with Twitter operations. At the end of 2022, he stopped using the datacenters in the city of Sacramento, stating that the city in the California region of the United States was “possibly the worst place to have a datacenter” because of the high temperature.

However, according to The Information, after this event it seems that Musk has changed his mind about the Sacramento heat and Twitter’s datacenters. He signed a lease with NTT Data, the operator of one of these facilities, to occupy the former Twitter space. It is believed that Tesla’s plan is to use the datacenter for machine learning workloads, including autonomous car simulations.

Tesla is also considering subletting a second datacenter in Sacramento, managed by Prime Data Centers, where Twitter still has a rental agreement, although it doesn’t use the facility.

In 2021, Tesla opened a datacenter in Shanghai, China, to handle the data collected from its vehicles in the country. It continues to guarantee that Chinese data is being managed in the country, even after an incident in mid-August at an airport in Hunan province that barred Tesla vehicles from the car park.

Airport staff said that a Tesla security feature that scans the car’s surroundings for suspicious activity appears to have put confidential data at risk. A video of the car park is circulating on the internet. Tesla responded that all surveillance data collected by the cars is stored on USB drives in the vehicle and cannot be accessed remotely. Tesla stores user data in its datacenter in China.

In another initiative around datacenters at the beginning of the year, in March 2023, Musk was in Mexico, where the development of new datacenters is gaining momentum, especially in Querétaro, because there is less risk of earthquakes and there is still land available to build the facilities. In addition, Mexico is a country with an automotive vocation in Latin America. Therefore, by bringing together automotive and datacenter potential in one country, it caught the attention of Musk, who announced plans to build Tesla’s largest factory in Mexico.

Why invest in data centers?

Speculation surrounding this recently opened datacenter position highlights the fact that Tesla invests in various technologies with huge data processing requirements. The main one is its electric vehicles full of sensors and processors that work with telemetry data in real time, for example.

“Musk’s company, which electrifies and automates vehicles as well as its production processes, is increasing its demand for datacenters, as it needs these infrastructures to store the data from its vehicles and operations,” says Holly Garcia, vice president of Panduit’s datacenter business unit.

Tesla is also working on the development of Dojo, a supercomputer aimed at training autonomous cars, based on four technological pillars: large volumes of real data, neural network training, software and hardware for vehicles. We can say that three of them are well known to datacenter environments with large processing capacities (big data, neural networks and software for autonomous cars).