GM Unveils AI Roadmap for Faster Vehicle Development — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- GM unveiled its “intelligence-defined vehicle” roadmap, combining AI, centralized computing and virtualized engineering to target a development cycle of just over two years instead of the typical four to five.
- Sterling Anderson said he does not think humanoid robots should be building cars.
- Sterling Anderson said GM can now launch thousands of virtual tests with much quicker feedback than the six months previously required for physical crash-test prototypes, giving engineers broader safety coverage without sacrificing development speed.
- GM co-trains transformer-based foundation models with decades of internal design files and outside data, while safety-critical inference runs onboard to avoid latency and dropouts; noncritical OnStar AI interactions can use cloud round trips.
- GM is rolling out a common electrical system and centralized compute across its portfolio, letting software teams work as if the company had only two models in production rather than many varied deployment targets that lack a common baseline.
- T1-2 platform will underpin next-generation trucks across GM brands, with differentiation handled through aesthetics, features and interaction patterns while a single engineering organization focuses on the best suspension, powertrain and drive unit, then scales them across the portfolio.
- Sterling Anderson said GM is designing genuine autonomy for personal ownership rather than labeling assisted driving “eyes off,” expanding by road type with no safety-critical fallback to the driver and handoffs only after the vehicle reaches a safe place.
Why it matters: For GM, the concrete change is a four-to-five-year vehicle-development process targeted at just over two years, with thousands of virtual tests returning feedback much faster than the six months previously required for physical crash-test prototypes. That gives software teams one portfolio-wide baseline as GM competes with Tesla and Chinese EV makers amid softening U.S. EV demand.
Ask SkimNews
