Deutsche Bank examines why technology forecasts repeatedly fail and what that means for the AI doom debate. Experts often misjudge timing, adoption, and second-order effects; with over $5 trillion of hyperscaler investment planned over five years, real-world outcomes may hinge more on enterprise integration than model capability.
AI doom: A brief history of bad tech predictions
Deutsche Bank
Adrian Cox
Research
10 Pages
Key Takeaways
Experts miss adoption: Hinton expected AI to outperform radiologists within 5–10 years; radiologist headcount instead rose about 10% over the past decade.
Planning fails often: Among 3,000-plus megaprojects, only 25% met budget, 2.8% met budget and schedule, and 0.2% also delivered expected benefits.
AI improves forecasting: AI weather models generate state-of-the-art forecasts 100,000 times faster and 10,000 times more energy-efficiently, extending predictability only from eight to nine days.
Join our newsletter to have all of this content + Exclusive Newsletter Bonus Content delivered to your inbox every week
