GM’s New EV Platform Uses AI and Compute to Compete

GM’s 2028 EV platform centers on software and computing GM’s fourth-generation EV platform signals a shift in what the automaker considers central to vehicle design. Rather than emphasizing traditional specifications such as suspension geometry, dash-to-axle ratios, or torsional rigidity, the announcement focused on battery chemistry, redesigned electric motors, and a new central computing platform and electrical architecture. The architecture is designed to connect systems throughout the car, including acceleration and power, steering, braking, active safety, and software. The emphasis reflects an industry in which ride and handling remain important, but software increasingly helps distinguish one vehicle from another. GM has described its goal as making vehicles more capable and helpful in everyday life. The company says its overhaul of the electrical and computational systems is intended to support faster software and more capable automated-driving features. The new platform is slated to debut in 2028 in the Cadillac Escalade IQ, which GM describes as a technological flagship. The automaker has also said an autonomous highway-driving feature is planned for that year, allowing drivers to take their hands and eyes off the road without needing to be ready to take over. A central computer will power vehicle systems and autonomy GM plans to consolidate many vehicle functions in a central compute unit (CCU) that is 40 times more powerful than its existing systems. Rather than scattering smaller computer chips throughout the vehicle, the new architecture will connect systems through a central platform. The CCU will handle propulsion and chassis controls, along with perception and active safety. Bringing those functions together is also intended to help pave the way for more advanced autonomy features. The platform will connect vehicle systems including acceleration and power, steering and braking, as well as software. Eliminating smaller chips at the edges is expected to cut wiring by 30%, reducing weight and simplifying assembly. GM says the CCU can also be used in fossil-fuel models, though it is set to debut in the electric Cadillac Escalade IQ in 2028. Centralized computing is part of the broader overhaul of the vehicle’s electrical and computational architecture, which GM aims to use to deliver faster software and more capable automated-driving features. GM uses AI tools to speed up vehicle development GM wants to cut its traditional four-to-five-year vehicle development cycle in half, matching a faster pace as Chinese automakers have shortened their own development timelines to two to three years. The company says enterprise AI tools have helped speed work across every sector of vehicle development. One example involved a new lithium-manganese-rich battery pack. Before engineers assembled the first physical prototype, they evaluated 300,000 virtual designs. That process helped move the battery concept to an approved program in just one month. The example offers a glimpse of GM’s revised development strategy: use virtual testing and enterprise AI to move from design exploration to program approval faster. The company’s announcements on AI came alongside plans for new battery chemistries, electric motors, and vehicle computing, as GM works to accelerate development while preserving profitability amid upheaval in the automotive industry. Competition from Tesla and Chinese automakers raises the stakes GM’s software push comes as competitors raise expectations for what a vehicle can do. The automaker has invested heavily in software and autonomous-vehicle technology, replacing Apple CarPlay in nearly all its EVs with a Google-based infotainment system. Its capitalized software amortization expense reached $866 million in 2025, reflecting spending whose costs are spread over time. The system is capable, though it does not yet match standards set by Tesla and Rivian. Tesla has been pushing the boundaries of autonomy, helped by its development of a software-defined vehicle. Meanwhile, Chinese manufacturers have cut development cycles to two to three years, compared with the traditional four to five. Their factories can produce more vehicles than Chinese consumers can absorb, and subsidies and lower labor costs have helped drive low-cost exports into markets from Europe to South America. GM aims to match the faster development cadence and has outlined a 2028 highway-driving feature designed to let drivers take their hands and eyes off the road without needing to be ready to take over. The planned feature raises the stakes: GM must turn its software investments and new vehicle architecture into competitive autonomy while keeping pace with rivals’ speed and costs.