General Motors is framing its next technology reset as a two-front contest: beat Chinese automakers at speed, and make autonomy safer by skipping the murky middle ground of supervised driving assistance.

On Wednesday, GM unveiled a plan to rebuild future vehicles around centralized computing, artificial intelligence and autonomous driving. The company says the architecture will replace scattered electronic control units with a central computer offering 40 times more computing power while cutting wiring by 30%. It will also turn OnStar into a more proactive AI assistant, beginning with a software update for the 2027 Cadillac Lyriq.¹

The more consequential deadline is 2028. GM plans to introduce an “eyes-off” highway-driving system in the Cadillac Escalade IQ, then carry it into pickups, large SUVs and future EVs. Chief product officer Sterling Anderson says the company is deliberately avoiding assistance systems that still demand constant driver supervision. “They can be complacent, and we will be safe,” he said of occupants in GM’s proposed hands-free highway system.¹

That is a direct challenge to the model represented by Tesla’s Full Self-Driving software. Elon Musk has argued that FSD itself is a major product draw, writing: “People are buying FSD with a car attached, not the other way around.”Elon Musk (elonmusk)² GM’s position is that partial automation can encourage drivers to over-trust a system that may still need them to intervene; its longer-term “Mission Mode” vision is a vehicle that can run errands without an owner aboard.¹

GM also concedes that Chinese manufacturers retain a structural edge: tightly integrated suppliers let them source parts and launch models faster than Western rivals. Rather than replicate that ecosystem, GM says it will mine decades of engineering and manufacturing records, use virtual testing and run development tasks in parallel.³

The target is to compress GM’s traditional four-to-five-year vehicle-development cycle to roughly two years. In one example, engineers virtually evaluated more than 300,000 designs for lithium manganese-rich batteries before building a physical prototype, moving from an initial pack concept to approval in one month.³ “With scale comes data,” Anderson said—a concise statement of GM’s wager that its old industrial heft can become an AI-era advantage.³