technologyScore 35/100Research

Transformer-based large language models driving unprecedented GPU demand; Hopper designed specifically for this workload

Jensen Huang· NVIDIA· AI· 2022-08-25· about NVIDIA (NVDA)
Hopper is a giant new generation because it is designed to perform this new type of AI model called Transformers... Hopper has an engine inside it called Transformer engine, with numerical formats and pipelines that allows us to do a spectacular job on Transformer-type of models, which includes large language models... Nobody could have predicted the profound importance of large language models. Large language models, excitement, innovation, ideas, companies, start-ups, industries, all exceeding everyone's expectations.

Why it matters

LLMs are driving a new class of compute demand that was not predicted; Hopper's Transformer engine is purpose-built for this. This signals sustained and accelerating demand for AI infrastructure beyond original hyperscaler projections.

Investment implication

Companies supporting LLM training and inference (data center operators, cloud providers, software optimization vendors) will see sustained capex cycles. Memory, interconnect, and storage vendors will benefit from 100B-1T parameter model training requirements.

Source

NVIDIA Q2 2023 Earnings Call (The Motley Fool)
← All signals