Nvidia says it can shrink LLM memory 20x without changing model weights

Nvidia's KV Cache Transform Coding (KVTC) compresses LLM key-value cache by 20x without model changes, cutting GPU memory costs and time-to-first-token by up to 8x for multi-turn AI applications.

MIT Finds Way To Shrink AI Memory 50x Without Losing Accuracy

This breakthrough could make AI far more practical for large-scale use as the method promises to cut cloud computing costs and process huge datasets faster.
VentureBeat

Chinese researchers unveil MemOS, the first 'memory operating system' that gives AI human-like recall

A team of researchers from leading institutions including Shanghai Jiao Tong University and Zhejiang University has developed what they're calling the first "memory operating system" for ai, ...
PCMag

RAM Reality Check: How Much Memory Does Your PC Actually Need in 2026?

The AI hardware boom is sending memory prices sky-high, so knowing exactly how much you need is more critical than ever. I've worked out the most realistic RAM goals for every type of PC.
Morningstar

Vectorize Breaks 90% on LongMemEval with Open-Source AI Agent Memory System

BOULDER, Colo., Dec. 16, 2025 /PRNewswire/ -- Vectorize today released Hindsight, an open-source memory system for AI agents that, for the first time, surpasses 90% accuracy on LongMemEval, the ...