Nvidia details efficiency of the NVFP4 format for LLM training — new paper reveals how NVFP4 offers benefits over FP8 and BF16

Nvidia details efficiency of the NVFP4 format for LLM training — new paper reveals how NVFP4 offers benefits over FP8 and BF16

When Nvidia began to disclose details about its new 4-bit floating point format — NVFP4 — earlier this year, it stated that while it is mainly designed for inference, it could also be used for AI training without significant loss in accuracy. Recently, the company released a paper describing how it managed to train a…

Read More
AI data centers are swallowing the world’s memory and storage supply, setting the stage for a pricing apocalypse that could last a decade

AI data centers are swallowing the world’s memory and storage supply, setting the stage for a pricing apocalypse that could last a decade

This free-to-access article was made possible by Tom’s Hardware Premium, where you can find in-depth news analysis, features and access to Bench. Nearly every analyst firm and memory maker is now warning of looming shortages of NAND and DRAM that will result in skyrocketing pricing for SSDs and memory over the coming months and years,…

Read More
Amid Intel’s deals, Intel Foundry Services remains notably absent — 18A and 14A are on the way, but success isn’t guaranteed

Amid Intel’s deals, Intel Foundry Services remains notably absent — 18A and 14A are on the way, but success isn’t guaranteed

It has been a busy few weeks for semiconductors. Apple debuted its next-generation silicon, Nvidia launched an unprecedented partnership, and TSMC denied rumors of a major tie-up with Intel. But across all these developments, one name has remained conspicuously absent: Intel Foundry Services. Take the biggest stories of the month. Apple’s new A19 and A19…

Read More
OpenAI might be building its own chip, but it’ll still be dependent on Nvidia — custom chip developed with Broadcom reportedly slips to Q3 2026

OpenAI might be building its own chip, but it’ll still be dependent on Nvidia — custom chip developed with Broadcom reportedly slips to Q3 2026

OpenAI’s long-rumored $10 billion partnership with Broadcom is already showing cracks. The company is widely understood to be developing a custom chip designed specifically for OpenAI’s inference workloads, but according to individuals familiar with the matter, the project has “hit snags”: OpenAI wanted more power, sooner, than Broadcom could deliver, and an internal push to…

Read More