1. Home ›
  2. Blog ›
  3. Why AI Is Not a Bubble: The Scaling Evidence Every Executive Needs to See

Why AI Is Not a Bubble: The Scaling Evidence Every Executive Needs to See

By Juan Beltrán · 2026-03-02

AI task complexity is doubling every 3 months, six times faster than Moore's Law. AI scores 90% on PhD-level benchmarks where human experts average 65%. AGI timelines have collapsed from 50 years to possibly 5. The bubble narrative is not just wrong. It is dangerously complacent.

Key takeaways

  • AI task complexity is doubling every 3 months, a rate 6x faster than the original Moore's Law. The tools your team evaluated six months ago are already obsolete
  • Every predicted scaling wall since 2019 has been broken by a new paradigm. What looks like a plateau is just the transition between stacked S-curves
  • AI now scores 90% on PhD-level science benchmarks where human experts average 65%, and has discovered entirely new physics laws
  • AGI timeline estimates have collapsed from 50 years away (2020) to a 25% chance by 2027 and 50% by 2031, according to surveyed AI researchers
  • Even the architects of this technology assign a 25% probability to catastrophic outcomes, making responsible governance not optional but existential

About the author

Juan Beltrán writes about AI transformation, CRM, data analytics and digital growth for enterprise leaders in complex B2B industries. Head of Digital Marketing, ABB Energy Industries. 17+ years in enterprise transformation. Based in Zug, Switzerland.

Disclaimer

This is a personal website. The views and opinions expressed here are my own and do not represent ABB or any current or former employer. All content is based on public information, personal experience and general professional knowledge. No confidential, proprietary, client-specific or employer-specific information is shared.

Canonical URL: https://juanbeltran.ch/blog/why-ai-is-not-a-bubble-scaling-evidence-executive-guide-2026