# Google's TurboQuant AI Compression Algorithm Wiped $100 Billion From Memory Chip Stocks This Week

_Friday, March 27, 2026 at 11:33 PM EDT · Infrastructure · Developing · Tier 1 — Major_

![Google's TurboQuant AI Compression Algorithm Wiped $100 Billion From Memory Chip Stocks This Week — Primary](https://images.ft.com/v3/image/raw/https%3A%2F%2Fd1e00ek4ebabms.cloudfront.net%2Fproduction%2Fde0a5310-e922-4dbc-9d31-7b6eabd64f83.jpg?source=next-article&fit=scale-down&quality=highest&width=700&dpr=1)

Google Research's TurboQuant compression algorithm sent US memory chip stocks down sharply this week, erasing roughly $100 billion in market value. Led by Micron's 15% drop, the selloff reflects a recalibration: if AI models can run efficiently on less memory, demand projections for HBM and DRAM must be revised downward. It is an early test of whether the AI memory supercycle is as durable as Wall Street assumed.

[Instagram]
A single Google research paper erased $100 billion from the US memory chip sector in one week. Google's TurboQuant compression algorithm reduces the amount of memory AI models need to run. Micron fell 15%. The market is repricing how much memory the AI boom actually requires.

[LinkedIn]
Google Research's TurboQuant compression algorithm sent US memory chip stocks down sharply this week, erasing roughly $100 billion in market value. Led by Micron's 15% drop, the selloff reflects a recalibration: if AI models can run efficiently on less memory, demand projections for HBM and DRAM must be revised downward. It is an early test of whether the AI memory supercycle is as durable as Wall Street assumed.

## Sources


---
Canonical: https://techandbusiness.org/newswire/newsDraft.b4bac002-3ccd-4278-b499-1e14245df9c7
Retrieved: 2026-07-21T23:47:03.782Z
Publisher: Tech & Business (techandbusiness.org)
