21 October 2026
GM wants to build a new car in two years, and it is betting on AI to get there
At an event in New York, General Motors set a goal of halving the time it takes to develop a vehicle, and showed off a new in-car computer and its own OnStar AI assistant.
Designing a new car usually takes the big carmakers four to five years. General Motors says it wants to do it in about two. At an event called GM Forward in Manhattan on Wednesday, the Detroit company explained how: more artificial intelligence in the engineering lab, far more computing power inside its vehicles, and an AI assistant it built itself.
What happened
The headline target is speed. GM is aiming for roughly two years from first concept to the start of production, though it gave no date for when it expects to get there.
Much of the time saving is meant to come from testing on computers before building anything. GM said it ran more than 300,000 virtual battery designs using its new lithium manganese rich (LMR) chemistry before it made a single physical prototype. Chief product officer Sterling Anderson said LMR will be the only battery type in GM's fourth generation of electric vehicles (EVs), and argued that a century of engineering records gives GM an edge. "In the age of AI, data is currency," he said.
The other pieces:
- A new central computer for both EVs and gas models, with 40 times the processing power of older systems and up to 50% faster cellular links. It arrives first on the $130,000 Cadillac Escalade IQ in 2028.
- Next-generation EVs that GM expects to be 25% more aerodynamic and 15% lighter, with smaller motors and battery packs.
- A home-grown OnStar AI assistant, sold by subscription, coming in the next few months to the 2027 Cadillac Lyriq through an over-the-air (wireless) software update. GM uses Google's Gemini for this today.
- Eyes-off highway driving, where the driver can stop watching the road, still planned for 2028, starting with the Escalade IQ.
Why it matters
The pressure is coming from China. Consulting firm AlixPartners estimates Chinese carmakers get vehicles to market twice as fast as traditional rivals, with 40% to 50% less investment and a 30% cost advantage. In the US, GM is still catching up with Tesla on wireless updates and driver-assistance features.
Speed matters more when so much of a car's value lives in software: a model that takes five years to design risks feeling dated by the time it reaches showrooms.
For scale, GM took in about $186 billion of revenue over the twelve months to June, according to its filings, and the whole company was valued at about $73 billion on the stock market as of October 19. It has also been five years since CEO Mary Barra said personal self-driving cars would arrive this decade, so the 2028 date has a clock on it.
What to watch
- Whether GM attaches a timeline to the two-year development goal.
- The OnStar AI launch on the Lyriq, and how many owners pay for it.
- The 2028 Escalade IQ, which is due to carry both the new computer and eyes-off driving.
GM's revenue, profit and debt history are on its Noorult company page.
The companies in this story
Based on reporting by CNBC: GM aims to shorten vehicle development time with AI, software push. Facts, not advice.
More news
- Anthropic's IPO, in numbers: 825 million shares at $70 to $78, and a $1.4 trillion price tag
- Trump says the Nvidia scandal didn't endanger US security, and stands by Lutnick
- Nvidia's first earnings without Huang: what Kress will be asked on Tuesday
- Buffett gives away $5.5 billion of Berkshire stock in his Thanksgiving letter
- Banks are using AI to trim back offices. Proof that it adds revenue is harder to find
- IMF chief says the world economy leans too hard on AI as circular deals top $1 trillion