Morgan Stanley warns clients that AI models will reach a critical self-improvement threshold between April and June 2026, triggering exponential capability increases that markets aren't pricing in. The bank's prediction follows warnings from AI executives at its recent TMT conference and aligns with OpenAI CEO Sam Altman's February statement that "extremely capable models" will arrive sooner than expected with a "faster takeoff" than originally anticipated.
Why it matters
Industries vulnerable to AI disruption—including wealth management, insurance, shipping, gaming, property management, and software—already saw stock declines earlier in 2026 and face accelerating price deflation and workforce reductions within months. The non-linear growth curve means enterprises have a compressed window to adapt before AI capabilities fundamentally reshape competitive dynamics and service pricing across knowledge work sectors.
What to do
Accelerate infrastructure investments for AI deployment now, before the April-June capability jump. Simultaneously identify and protect business lines that require irreplaceable human judgment or relationships, as these assets will command premium valuations as AI commoditizes routine services.