Archive · 25 Sep 2025
TSMC Uses AI for More Energy-Efficient AI Chips
Imagine computer chips, like tiny brains, that use a lot of power.
TSMC is using special AI programs to design these chips to use less energy.
These AI programs can help engineers find better chip designs much faster than humans.
The new chips use multiple smaller chips working together.
This allows them to work better together, using less power.
TSMC is using AI to design more energy-efficient computing chips.
The goal is to increase energy efficiency by roughly 10 times.
Chip design firms like Cadence and Synopsys provide AI software.
AI software solutions found better chip designs and faster than human engineers.
New technologies like optical connections are being explored to improve data transfer.
- Who
- TSMC, Nvidia, Cadence, Synopsys, Meta Platforms
- What
- Using AI to design more energy-efficient chips
- When
- Wednesday
- Where
- Silicon Valley
- Why
- To reduce energy consumption of AI chips
- Key Players
- TSMC, Nvidia, Cadence, Synopsys, Meta Platforms
- Main Goal
- Improve energy efficiency of AI chips
- Technology
- AI-powered software
- Target Improvement
- About 10x energy efficiency boost
- Chip Design
- Multiple 'chiplets' packaged together
- Conference Location
- Silicon Valley
- Data Transfer Tech
- Optical connections being explored
