There’s a paradox playing out in the AI industry: funding is at all-time highs, but layoffs are accelerating. According to layoffs.fyi, over 15,000 employees at AI companies have been laid off in 2025 — a 200% increase from the same period in 2024.
The companies affected span every stage, from Series A startups like Adept AI (which laid off 30% of staff before being acquired by Amazon) to later-stage companies like Stability AI, Inflection AI, and Hugging Face (which reduced headcount by 10% despite raising $235 million).
What’s Happening
The common thread: AI capabilities are advancing faster than companies can build products around them. A team of 5 engineers using GPT-5 can now build what required 50 engineers using GPT-3.5. Companies that raised money to hire large teams are discovering they don’t need them.
“Every time a new model drops, half my product roadmap becomes trivially implementable,” said the CEO of one Series B AI startup who spoke on condition of anonymity. “I raised money to hire 80 people. Now I need 20.”
The Structural Shift
This isn’t a downturn — it’s a structural realignment. AI is making AI development itself more efficient. The companies that survive will be small, capital-efficient teams that leverage frontier models as infrastructure rather than trying to compete with them.
What It Means for Job Seekers
The traditional AI engineer role — training models, building ML pipelines — is being commoditized. The roles growing in demand: AI product managers who understand what to build, domain experts who can evaluate AI outputs, and “AI operations” engineers who keep production systems reliable.