Newshot

Former Intel CEO Pat Gelsinger: Incremental chip performance gains can't keep up as AI demands a "10,000x efficiency leap"

AI summary

Former Intel CEO Pat Gelsinger said that as AI requires a "10,000x efficiency leap," incremental chip improvements will not be enough, while denying that Moore's Law has come to an end. He argued that AI needs a radical 10,000x jump in computing efficiency rather than routine generational upgrades, and that reviving Moore's Law depends on fundamental breakthroughs in five core engineering areas: optics, power delivery, thermal management, materials, and packaging.

Translation(model translation, not the source's own text)

Former Intel CEO Gelsinger: As AI Demands a "10,000x Efficiency Leap," Incremental Chip Gains Won't Be Enough

IT之家 November 21 news, according to Wccftech, former Intel CEO Pat Gelsinger said that as artificial intelligence demands a "10,000x efficiency leap," incremental chip gains will not be enough, while he rejected the notion that Moore's Law has come to an end.

At the 2026 COSM technology summit, Pat Gelsinger took part in a fireside chat titled "Reality Check: Beyond Chips."

During the conversation, Gelsinger pushed back on the view that semiconductor scaling has reached its physical limits, arguing that AI needs not routine generational upgrades but a radical 10,000x leap in computing efficiency. He also noted that whether Moore's Law can be revived depends on fundamental breakthroughs in five core engineering areas (IT之家 note: optics, power delivery, thermal management, materials, and packaging).

After the summit, Pat Gelsinger wrote on X: "Moore's Law is dead? I say it's time to wake it up!" Although some industry observers may argue that traditional Moore's Law scaling is over, Gelsinger insists that semiconductor engineering is entering a new era of multidisciplinary innovation, saying: "After more than 40 years in the semiconductor industry, I believe our greatest innovations are still ahead!"

Since joining the venture capital firm Playground Global as a general partner in March 2025, Gelsinger has been actively backing startups targeting these hardware bottlenecks, including the U.S. government-backed lithography startup xLight (which aims to break through ASML's 13.5-nanometer extreme ultraviolet EUV wavelength barrier), as well as nuclear energy capacity company Alva Energy.

Beyond lithography and power generation, Gelsinger's technology roadmap also targets optical interconnect and silicon photonics, aiming to replace traditional copper wiring with optical signals for data transmission.

  • "Former Intel CEO Gelsinger Calls HBM 'Terrible Memory'"
  • "Former Intel CEO Gelsinger Bets on New AI Memory: Z2RAM Aims to Bypass EUV Lithography, Sampling Within 4 Years"

Related Articles

  • Former Intel CEO Gelsinger Reflects: I Really Did Look Down on Nvidia GPUs Back Then
  • Former Intel CEO Pat Gelsinger Critiques His Old Employer: Several of His Predecessors Lacked Professional Backgrounds and Weren't Fit to Lead the Company
  • Former Intel CEO Gelsinger Sounds Off: Wall Street Doesn't Understand Long-Term Investment, Chip Fabs Take a Decade to Pay Back
  • Former Intel CEO Gelsinger: Resignation Was a Decision Made by a "Third Party," No Chance to Finish Unfinished Business
  • Former Intel CEO Gelsinger to Receive $7.85 Million Severance
  • Intel CEO Gelsinger: Don't Listen to Jensen's Nonsense, Moore's Law Is Still Valid

The translation is machine-generated and individual terms may be off. Nothing is rewritten or cut; the original is authoritative.

Published by IT之家
This site only summarises and aggregates; copyright stays with the original author. Headlines and summaries follow the source — where they differ, the original wins.
Read the original ↗