Automakers are rewriting their org charts around AI, and the bodies are piling up. TechCrunch AI reports that General Motors just cut more than 10% of its IT department, about 600 salaried employees, in what the company calls a “deliberate skills swap.” GM is hiring back, but not for the roles it eliminated. The new shopping list: AI-native developers, data engineers, cloud architects, agent and model builders, and prompt engineers.
This isn’t a one-for-one trade. It’s a net loss in headcount, and it’s happening across Detroit.
The numbers behind the shift
Ford, GM, and Stellantis have collectively cut more than 20,000 U.S. salaried jobs this decade, according to CNBC figures cited by TechCrunch AI. That’s 19% of their combined white-collar workforce gone from recent peaks. Not all of it is AI-driven, but the technology is the through-line.
What GM wants now is telling. The company isn’t looking for people who use AI as a productivity boost. It wants people who can build with AI from the ground up: design the systems, train the models, engineer the pipelines. The bar moved.
Why this matters beyond automotive
Auto is a leading indicator. When a 120-year-old industry with strong unions and conservative HR practices starts swapping IT generalists for AI specialists at this pace, the rest of the economy is paying attention.
Three dynamics worth watching:
- The skills gap is widening fast. Companies aren’t waiting for retraining programs to catch up. They’re cutting and rehiring in the same quarter.
- “AI-native” is the new credential. Knowing how to use ChatGPT won’t save your role. Knowing how to ship agentic systems will.
- Some companies don’t know what they’re doing yet. TechCrunch AI notes that engineers and founders report many of these AI bets are speculative. Layoffs are happening anyway.
The bright spot: real revenue from real data
Samsara is the counterexample worth studying. The fleet-monitoring company spent a decade collecting camera footage from millions of trucks. It then trained a model to spot potholes and predict how fast they’ll deteriorate. Now it’s selling that product to cities, with Chicago already under contract.
That’s the playbook: own proprietary data, train a vertical model, sell it into a market that didn’t exist before. Most companies talking about AI strategy don’t have either piece in place.
What practitioners should do now
If you work in IT, engineering, or any white-collar role adjacent to software:
- Audit your AI fluency honestly. Can you build, or just prompt? The market is paying for the first.
- Pick a vertical. Generic AI skills are commoditizing. Domain-specific AI work (automotive, healthcare, logistics) commands premiums.
- Get hands-on with agents and pipelines. These are the specific phrases showing up in job descriptions right now.
For business leaders: the GM approach is going to spread. Cut deeply, rehire selectively, accept the net-negative headcount as the cost of the transition. The companies waiting for a graceful path won’t find one.
The forecast
Expect more Fortune 500 names to announce “skills swaps” in the next 12 months. Expect the term itself to become a euphemism that fools nobody. And expect the wage gap between AI-native talent and everyone else to keep widening, because the supply of people who can actually build production AI systems isn’t catching up to demand.
The automakers are showing the rest of the economy what the next two years look like. Full details at the original TechCrunch AI report.