# AI and Semiconductor Memory: More, More, More

- Show: [EE Times Current](https://hardwarepodcasts.com/podcast/ee-times-current)
- Published: 2021-10-01
- Length: 40m
- Page: https://hardwarepodcasts.com/podcast/ee-times-current/2021-10-01-ai-and-semiconductor-memory-more-more-more
- Listen: https://podcast.show/aspencore/episode/81562351/

## Summary

Steven Woo, a Rambus fellow and distinguished inventor, explains how AI's exponential growth in model size and training data is creating unprecedented demands on memory capacity, bandwidth, and power efficiency. He details the industry's shift toward High Bandwidth Memory (HBM) and GDDR to handle these loads, while noting that emerging memories like ReRAM and MRAM face reliability hurdles. Woo also discusses how architectural innovations like Compute Express Link (CXL) and chip stacking are critical for reducing the power and latency associated with moving data between processors and memory.

## Chapters

- [0:00](https://hardwarepodcasts.com/podcast/ee-times-current/2021-10-01-ai-and-semiconductor-memory-more-more-more#t=0) Introduction: AI's Impact on Semiconductor Memory
- [1:43](https://hardwarepodcasts.com/podcast/ee-times-current/2021-10-01-ai-and-semiconductor-memory-more-more-more#t=103) The Shift from Performance-Driven to Memory-Driven Design
- [11:09](https://hardwarepodcasts.com/podcast/ee-times-current/2021-10-01-ai-and-semiconductor-memory-more-more-more#t=669) AI's Demand for Capacity, Bandwidth, and Power Efficiency
- [15:44](https://hardwarepodcasts.com/podcast/ee-times-current/2021-10-01-ai-and-semiconductor-memory-more-more-more#t=944) HBM vs. GDDR and the Status of Emerging Memories
- [22:29](https://hardwarepodcasts.com/podcast/ee-times-current/2021-10-01-ai-and-semiconductor-memory-more-more-more#t=1349) Moore's Law, Data Movement, and Chip Stacking
- [27:14](https://hardwarepodcasts.com/podcast/ee-times-current/2021-10-01-ai-and-semiconductor-memory-more-more-more#t=1634) Compute Express Link (CXL) and System Disaggregation
- [33:38](https://hardwarepodcasts.com/podcast/ee-times-current/2021-10-01-ai-and-semiconductor-memory-more-more-more#t=2018) Sparse Models and Multi-Chip Communication

## Topics

- [PCB design & layout](https://hardwarepodcasts.com/topic/pcb-design)
- [Chips, FPGAs & semiconductors](https://hardwarepodcasts.com/topic/chips-and-fpgas)
- [Scaling production](https://hardwarepodcasts.com/topic/scaling-production)
- [Mechanical engineering](https://hardwarepodcasts.com/topic/mechanical-engineering)
- [AI & machine learning](https://hardwarepodcasts.com/topic/ai-and-machine-learning)
- [Batteries & low-power design](https://hardwarepodcasts.com/topic/batteries-and-low-power)

## Most relevant to

large enterprises

## Mentioned

- Rambus (company)
- NVIDIA (company)
- HBM (product)
- GDDR (product)
- CXL (product)
- Google TPU (product)
- Cerebras (company)
- Graphcore (company)
- Intel (company)

## Related episodes from other shows

- [EP 163: Breaking the Memory Wall: Micron’s Strategy for the AI Era](https://hardwarepodcasts.com/podcast/the-circuit/2026-05-05-ep-163-breaking-the-memory-wall-microns-strategy-for-the-ai-era) (The Circuit)
- [Episode 36: The Wide World of Memory](https://hardwarepodcasts.com/podcast/the-circuit/2023-10-02-episode-36-the-wide-world-of-memory) (The Circuit)
- [EP 187: Inside the Future of Memory Summit & AMD’s "Boring" Bullishness](https://hardwarepodcasts.com/podcast/the-circuit/2026-08-09-ep-187-inside-the-future-of-memory-summit-amds-boring-bullishness) (The Circuit)
- [EP 141: Talking Bubbles, ARM and QCOM Earnings, Memory/Storage Cyclicality](https://hardwarepodcasts.com/podcast/the-circuit/2025-11-09-ep-141-talking-bubbles-arm-and-qcom-earnings-memory-storage-cyclicality) (The Circuit)
- [Memory Pricing Pressures, Meta's Muse AI, and the Agentic CPU Crunch](https://hardwarepodcasts.com/podcast/the-circuit/2026-09-28-memory-pricing-pressures-metas-muse-ai-and-the-agentic-cpu-crunch) (The Circuit)

## Citing this page

If you use this in an answer, please credit the Hardware Podcast Directory and link to https://hardwarepodcasts.com/podcast/ee-times-current/2021-10-01-ai-and-semiconductor-memory-more-more-more. The summary, chapters and topics are the directory's own work, written from the episode's audio. The episode itself, its title and its audio belong to EE Times Current.
