General Motors Is Cutting Its Development Cycles in Half
Summary
General Motors is dramatically accelerating its vehicle development cycles, aiming to match the two-year or less timeframe achieved by Chinese automakers like BYD. Led by Chief Product Officer Sterling Anderson, GM is leveraging AI and advanced simulation tools to integrate design and engineering processes virtually. This "third epoch" of design allows engineers to simultaneously develop and optimize hardware and software, perform virtual calibrations, and run thousands of simulated scenarios—including crash tests and driver-in-the-loop simulations—before physical prototyping. This approach halved the development time for the electric GMC Hummer, reducing it from four to five years to just two. GM is applying these methods across various programs, including self-driving cars, LMR batteries, Cadillac's Formula 1 racing, military defense systems, Lunar Outpost's Pegasus rover, and even using AI for generative design on components like the Chevrolet Corvette's hatch support brackets.
Key takeaway
For Directors of AI/ML overseeing product development, GM's success demonstrates that integrating AI and simulation early can drastically cut development cycles. You should prioritize investing in virtual integration platforms and physics-based modeling to enable simultaneous hardware/software optimization. This approach allows your teams to perform comprehensive virtual calibrations and scenario testing, hardening designs against real-world conditions before physical prototyping, thereby accelerating market readiness and reducing costs.
Key insights
AI and simulation enable simultaneous virtual development, collapsing design cycles and enhancing product robustness.
Principles
- Virtual integration collapses design silos.
- Physics-based models harden designs.
- AI accelerates complex simulations.
Method
GM's method involves a proprietary virtual environment for simultaneous hardware/software development, physics-based simulations across thousands of scenarios, and AI-powered generative design for components.
In practice
- Simulate 100 days of driving in one day.
- Cut crash simulation from 15 hours to 1 minute.
- Use generative AI for lighter, stiffer parts.
Topics
- Automotive Development
- AI in Engineering
- Simulation Technology
- Generative Design
- Virtual Integration
- Product Lifecycle Management
Best for: AI Product Manager, CTO, VP of Engineering/Data, AI Engineer, Director of AI/ML, Research Scientist
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by IEEE Spectrum.