This Former Intel CEO Wants to Jumpstart Moore’s Law With Light
Summary
Former Intel CEO Pat Gelsinger, now a general partner at deep tech venture firm Playground Capital, aims to reawaken Moore's Law by advancing semiconductor lithography. He believes overcoming the physical and cost limits of shrinking transistors requires breakthroughs in etching chips with light, moving beyond ASML's current 13.5-nanometer wavelength technology. Gelsinger serves on the board of xLight, a company developing novel lithography techniques, including free-electron lasers for smaller 5-, 4-, 3-, and 2-nanometer light, intending to enhance existing ASML machines. The semiconductor industry, projected to reach \$1 trillion by next year, is heavily influenced by AI, which Gelsinger believes demands new chip architectures focused on inference, not just GPUs, to achieve 10,000x improvements. He also highlights energy capacity as a critical AI bottleneck, supporting companies like Alva Energy for nuclear upgrades, and emphasizes the need for robust industry or government-led integrity and security reviews for rapidly evolving foundational AI models.
Key takeaway
For investors evaluating deep tech or Directors of AI/ML planning future infrastructure, recognize that current GPU-centric approaches and energy limitations are significant bottlenecks. Your long-term AI strategy should prioritize investments in advanced lithography, novel inference-optimized chip architectures, and stacked memory solutions to achieve substantial performance gains. Additionally, advocate for energy capacity expansion and scrutinize AI model integrity and security, as these foundational elements will dictate the economic viability and ethical deployment of future AI systems.
Key insights
Reviving Moore's Law and advancing AI hinges on lithography breakthroughs and expanding energy infrastructure.
Principles
- Deep tech investing demands rigorous technical diligence.
- Lithography remains core to semiconductor progress.
- Energy capacity is economic capacity in the AI age.
Method
Assess deep tech founders via a deeply technical team, rigorous diligence, interviews, and background checks, focusing on the "best team" solving a hard problem.
In practice
- Invest in advanced lithography beyond 13.5nm.
- Develop stacked memory architectures for AI.
- Support nuclear energy upgrades for AI capacity.
Topics
- Moore's Law
- Semiconductor Lithography
- Deep Tech Investment
- AI Hardware
- Energy Capacity
- AI Regulation
Best for: Investor, Director of AI/ML, Policy Maker
Related on AIssential
See Counsel's argued verdicts on the open AI decisions leaders are weighing →
Editorial summary, takeaway, and curation by AIssential. Original article published by WIRED - Ai.