Full Breakdown
GM Rolls Out Central Compute Architecture and AI-Powered OnStar Assistant to Accelerate EV and Autonomy Plans
By Drooid · · How we work
New Central Compute Platform and OnStar AI Announcement
At its GM Forward showcase in New York, General Motors introduced a fourth-generation electric-vehicle (EV) platform built around a single central compute unit (CCU). The CCU, described as 40 times more powerful than the company’s existing electronics, will manage propulsion, chassis, perception and active-safety functions across both electric and fossil-fuel models. The same event unveiled an AI-driven OnStar assistant that will debut in the 2027 Cadillac Lyriq via an over-the-air update, with additional models slated to follow.
Background & Context
The push reflects GM’s response to a software-first market led by Tesla, Rivian and rapidly iterating Chinese manufacturers that have shortened development cycles to two-to-three years. GM executives have framed software as the primary differentiator for modern vehicles, moving away from traditional mechanical specifications.
Data & Statistics
- GM reported a capitalized software amortization expense of $866 million in 2025.
- The new CCU is expected to reduce wiring by 30 percent, lowering vehicle weight and assembly complexity.
- Engineers simulated 300,000 virtual designs of a lithium-manganese-rich battery before building the first physical prototype.
- GM claims AI-enabled processes can shrink new-vehicle development from four-to-five years to roughly two years.
- The 2028 Cadillac Escalade IQ will be the first model to feature “eyes-off” autonomous highway driving, extending the existing Super Cruise system.
Official Statements & Responses
He added that the company’s decades of data support training its AI models.
Mary Barra, GM’s chief executive officer, described OnStar’s evolution as a shift from a call-center service to a “connected intelligence platform” that can anticipate driver needs and integrate calendar and weather data.
A GM spokesperson emphasized that data from the new Connected Cameras feature will be end-to-end encrypted, stored only for limited periods, and inaccessible to the company. The spokesperson also noted that any sharing of data will be “feature-specific and consent-based.”
Criticism & Opposition
Earlier in 2026, the Federal Trade Commission issued an order prohibiting GM and OnStar from sharing precise geolocation and driver-behavior data with consumer-reporting agencies for five years. The order requires explicit consumer consent for data collection and mandates transparent disclosures about data use. Critics argue that the rollout of AI-driven services could re-introduce privacy risks, especially given the subscription-based model for Connected Cameras and the new OnStar app.
Verbatim Quotes
- “OnStar is now transitioning to be a connected intelligence platform, not just the call center that it’s historically been known as, but a mechanism for interpreting, understanding, and providing predictive, proactive support to its owners,” — GM’s VP of global product and chief product officer Sterling Anderson
- “You can imagine what that does for iteration speed,” — Sterling Anderson, executive vice president, global product, and chief product officer at GM
- “Our vision is that any of our vehicles should drive itself anywhere,” — Anderson, GM spokesperson
- “We're evolving OnStar to become a true driving partner,” — GM CEO Mary Barra
What’s Next
- 2027: OnStar AI Assistant launches in the Cadillac Lyriq; a unified OnStar mobile app replaces brand-specific apps.
- 2028: Cadillac Escalade IQ introduces the CCU and “eyes-off” autonomous highway driving; the central architecture will later expand to full-size trucks, SUVs and other GM models.
- Ongoing: GM plans to extend AI-accelerated development methods to future EV platforms, aiming to maintain a competitive cadence against Chinese rivals.