GM has just mapped out its entire future. It works with AI, eyeless driving and a central computer

General Motors aims to send a message: we are a modern company that can compete with and outperform less experienced newcomers. The technology presented and discussed impressed, but the devil is in the details or, historically at GM, the execution, as AutoPacific director of industrial analysis Paul Waatti said. The reader.

In New York, GM’s leadership team of CEO Mary Barra, President Mark Reuss and Chief Product Officer Sterling Anderson outlined the automaker’s near- and long-term future. It’s all built around a new, centralized electrical architecture that will underpin everything else: AI-accelerated product development, a complete reboot of OnStar as the center of ownership and user experience, next-generation electric vehicles, and blind-driving autonomy that extends across the entire lineup, including the automaker’s trucks. GM is already implementing some of this internally. Owners will get their first taste of this in the coming months with wider deployments starting in 2028.

The use of AI for vehicle development is already underway within GM. Anderson explained that integrating AI would cut the current development time in half, from four to five years. Additionally, aerodynamic tests that used to take 60 hours can now take 60 seconds. Structural analysis that previously took months can now take hours. “You can imagine what that does to the speed of iteration,” Anderson said.

All of this is made possible, according to Anderson, by GM leveraging its historical reserves of data to power machine learning systems. “GM’s size and scale with its data is not an anchor, it’s an advantage. Decades of data and experience. Smaller players can build models, but can’t match GM’s data,” Anderson said.

All of this will save not only time but also money through efficiency and streamlining of design, engineering, testing and even manufacturing. When asked if those savings would trickle down to the consumer in future products, Barra dodged the question and noted that GM has a huge line of vehicles that it already sells today for $30,000 or less at a profit.

OnStar will become your AI assistant, both in the car and on your phone. At least that’s GM’s vision of your vehicle ownership experience. Chevrolet, Cadillac, Buick and GMC’s iOS and Android apps will be replaced next year with a new OnStar mobile phone app that will unify the experience and provide a single place for the entire family’s vehicles, regardless of the badge on the front. Parents will be able to track both vehicles and people through device permissions (such as the Find My integration on an iPhone). This experience will work with both gasoline and electric GM vehicles. And of course, AI will be built right into the app so a user can ask it to make sure the vehicle is started and warm 15 minutes before the schedule says it needs to leave.

Before the new OnStar app launches, a new OnStar AI in-car assistant will launch later this year. First appeared in the 2027 Lyriq and later in GM’s broader lineup across gasoline and electric vehicles. Like other automakers, the OnStar AI agent will be deeply integrated into the vehicle’s subsystems and the GM team is coding the software to ensure it can provide true understanding of the vehicle. Using a natural language structure, a driver will be able to say, “Hey OnStar, what was that vehicle warning or alert sound coming from the car earlier and what did it mean?” » The vehicle will explain this alert and/or noise from earlier. For generations who grew up with OnStar connecting you to a live human to talk with this feature will always exist.

Future vehicles, starting with the 2028 Cadillac Escalade IQ, will run on GM’s new electrical architecture that includes a central computing unit. In an age where modern cars can have more than 80 to 100 ECUs, delivering new features via over-the-air software updates is complicated. Rivian made news with its first generation vehicles and only 17 ECUs, then again by revealing the second generation R1 models which reduced the number to just seven ECUs. GM is now preparing to deliver vehicles equipped with a single central computer.

This new system delivers a truly software-defined vehicle, lower latency, faster operations, 30% less wiring and 40 times more power than current systems. A spokesperson said the new mainframe had the power of 40 laptops. The system will create room for future features and expand beyond the Cadillac Escalade IQ to the automaker’s full-size trucks and SUVs, both gas and electric.

Additionally, the 2028 Cadillac Escalade IQ will debut GM’s eyeless autonomous driving system. This will take today’s Super Cruise hands-free driver assistance system and make it so drivers don’t have to watch the road. You can watch a movie, read a book, write emails, or hold court with your family to explain why you don’t turn that car around even though you’re not even driving it.

All of these new systems and hardware elements lead to next-generation electric vehicles. GM isn’t committing to a timeline for when we’ll see its fourth-generation EVs, but has said they will be powered by LMR batteries with 33% higher energy density than LFP batteries (which GM doesn’t currently use in its EVs), new electric motors that have 40% fewer parts than today’s Ultium motors, and 20% more range per kWh. The new electric vehicles are expected to go further, weigh less and be more efficient than the current electric vehicles sold by GM.

GM was cautious about when we’ll see the big finale here, namely next-generation electric vehicles, but a spokesperson let slip that development is expected to begin in 2028, which, given the new accelerated 2-2.5 year development cycle, means we could see the first next-generation electric vehicles on the road around 2030.

Do you have a story about the upcoming product? Send us a message at tips@thedrive.com