/Patent Tracker
NvidiaPatent applicationAI & softwareInterest 6/10

Nvidia patents method to speed up AI text generation with parallel processing

Nvidia describes a technique where one group of neural networks starts generating a sequence of words while a second group prepares the next word in parallel. This overlapping process, called prefill, aims to reduce the waiting time between generating each part of the output.

Why it mattersThis could make AI responses feel faster and more responsive by reducing the idle time between processing steps.

In the filing’s own words

Apparatuses, systems, and techniques to cause one or more neural networks to perform one or more prefill operations on a token of a set of tokens before the set of tokens is completely generated. In at least one embodiment, a processor comprises circuitry to cause one or more first neural networks to begin generating a set of tokens and cause one or more second neural networks to perform one or more prefill operations on a token of the set of tokens before the set of tokens is completely generated.

What exactly they claim (claim 1)

1. One or more processors comprising processing circuitry to: cause one or more first neural networks to begin generating a set of tokens; and cause one or more second neural networks to perform one or more prefill operations on a first token of the set of tokens in parallel with the generation of a subsequent token of the set of tokens.

Facts

Company
Nvidia (NVIDIA Corporation)
Official title
GENERATING TOKENS USING ONE OR MORE NEURAL NETWORKS
Publication number
US20260278386A1
Application number
19180502
Filed
Apr 16, 2025
Published
Sep 17, 2026
Inventors
3
Claims
20

Classified under

G06N 3/0895 Physics & computing

More from Nvidia in AI & software

Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.