General Motors (GM) has promised to catch up with Tesla (TSLA) in electric vehicles by the end of the current decade.
But the legacy automaker is not behind its big rival in autonomous vehicles, another segment considered to be part of auto industry's future.
Indeed, its Cruise subsidiary, which specializes in developing autonomous technologies, even seems to have an edge because for a few months the company has been authorized to operate a robotaxi service in San Francisco.
These self-driving vehicles are allowed to carry people wherever they want to go in the city from 10 p.m. to 6 a.m. and only when there’s no "heavy rain, heavy fog, heavy smoke, hail, sleet, or snow.”
The rides are limited to "select streets in San Francisco,” the California Public Utilities Commission said last June. The cars involved can drive at a maximum speed of 30 mph.
In-House Chips
It's this advantage that GM doesn't want to lose as its rivals continue to refine their technologies.
For example, Tesla has just rolled out to 100,000 of its vehicle owners the latest version of Full-Self-Driving software, the advanced option of Autopilot, which is its driver-assistance system. Even though FSD enables the vehicles of Elon Musk's group to perform important maneuvers alone, the system does not make them autonomous.
GM via Cruise has just decided to start producing chips that power its autonomous vehicles. The goal is to reduce reliance on third-party suppliers and to lower costs at a time when shortages of semiconductors have forced automakers to temporarily suspend assembly of certain models.
Today, the waiting time between an order of semiconductors and the delivery has lengthened sharply. The problem is that autonomous vehicles that are loaded with technology require more chips than other vehicles.
"Two years ago, we were paying a lot of money for a [graphics-processing unit] from a famous vendor," said Carl Jenkins, vice president for hardware engineering at Cruise, in an interview with Reuters. He referenced Nvidia (NVDA), the prominent maker of GPUs.
"There is no negotiation because we’re tiny volume. We couldn’t negotiate at all. So that’s why I said, okay, then we have to take control of our own destiny,” he added.
Production Will Start in 2025
Unhappy with Nvidia's practices, Cruise has decided to develop its own chips, which will be used in Origin, its model that has no pedals and no steering wheel, expected in 2023.
Origin, which was presented in 2020, is designed as a shared vehicle. It will be rolling all the time, picking up and dropping off passengers at various destinations.
GM's decision puts the company in the driver's seat of its future in autonomous vehicles. The company also joins Tesla and Volkswagen (VLKAF) which have decided to reduce their reliance on semiconductor manufacturers by designing their own in-house chips.
Cruise has already developed four chips: a computing chip called Horta, which is the brain of the vehicle; a sensor data processor called Dune, a third unnamed chip that will process information from the radar chip, and an unspecified chip that will be identified later.
The firm started developing its in-house chips in 2020. The technology is still at an experimental stage but production will start in 2025.
The type of chips designed by Cruise use ARM processors. It's the same type of semiconductor used for phones and tablets because they conserve energy. The company assures that they will improve energy efficiency by lowering Origin's power consumption.
By developing in-house chips, Cruise hopes to gain control over its supply chain and lower the cost of self-driving vehicles. GM has never hidden its ambition to sell as many autonomous vehicles as possible before the end of the current decade.
