Get the newsletter
Nvidia · Chip · Filed Jun 3, 2026 · USPTO published Oct 1, 2026 · Oct 1 release

Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It

Every time an AI model trains on new data, that data has to be cleaned up and reformatted before the GPU can touch it. Nvidia thinks the network card should be doing that work, not the GPU.

A processing system with a data processing unit (DPU) and graphics processing unit (GPU) connected to external devices. Drawing from patent filing US 2026/0300779 A1.
A processing system with a data processing unit (DPU) and graphics processing unit (GPU) connected to external devices.All 48 drawings ↓
Publication number
US 2026/0300779 A1(US20260300779A1)
Applicant
NVIDIA Corporation
Filed · Published
Jun 3, 2026 · Oct 1, 2026
Inventors
Alvin Ihsani, Carl Everett Lacey Jr., Shaul Arazi, Elena Agostini, Penn Tasinga, Dana Groff, Dotan David Levi, Wojciech Wasko, Vishwesh Nath, Sachidanand Alle
Status when we published
Waiting for an examiner
Document
1 claim
US classification
382/248
Parent application
Continuation of 18/076,221 (filed Dec 6, 2022)(18076221)

What Nvidia's network-card preprocessing actually does

What happens when a data pipeline keeps your GPU sitting idle, waiting for work it could already be doing? That bottleneck is surprisingly common in AI data centers, and it eats into the performance you're paying for.

Nvidia's patent describes moving that prep work to the network interface card, the component that handles incoming data before it reaches anything else. Instead of handing raw data to the GPU and making it reformat everything, the network card converts it into the right shape, then parks it in shared memory where the GPU can grab it immediately.

The upside for you, if you're running large-scale AI workloads, is that your expensive GPU spends more time doing actual computation and less time doing housekeeping. It's a bit like having a prep cook chop vegetables so the head chef never has to stop cooking.

From the filing · CLAIM 1
… a network interface separate from the at least one accelerator to convert the data and store the data in the memory in preparation of performance of the at least one operation by the at least one accelerator.

Translation: A network chip prepares incoming data before handing it off to the main processor.

How the network interface converts and stores data for the GPU

The patent describes a system with two main pieces: at least one accelerator (in practice, a GPU or similar parallel processor) and a network interface that operates independently of that accelerator.

The network interface's new job is data conversion: taking data that arrives over a network in one format and converting it to a second format that the accelerator can process directly. Once converted, the network interface writes the data into shared memory that the accelerator can access.

The accelerator then picks up the already-formatted data and runs its computations, which could include AI training or inference operations, without needing to spend any of its own cycles on the reformatting step.

  • Data arrives at the network interface in its original format
  • The network interface converts it to a GPU-compatible format
  • Converted data is stored in memory accessible to the accelerator
  • The accelerator processes it without doing any format conversion itself

The core claim is that the network interface is separate from the accelerator, meaning it has its own processing resources dedicated to this task rather than borrowing compute from the GPU.

From the filing · THE ABSTRACT
… converting the data to a second format, storing the converted data in memory accessible by at least one parallel processing unit, and processing the converted data stored in the memory using the at least one parallel processing unit.

Translation: The system reformats the data and saves it in shared memory so the graphics chip can process it.

What this means for AI training speed and GPU efficiency

GPUs are expensive, and their time is valuable. Any task a GPU does that isn't matrix multiplication or model computation is, in a sense, wasted money. Offloading data preparation to a dedicated component lower in the stack is a well-understood design principle, but applying it specifically at the network interface level is a meaningful architectural choice for AI infrastructure.

Nvidia's interest in data-center pipeline efficiency shows up in multiple layers of their stack. For operators running large AI training clusters, reducing the preprocessing burden on accelerators could translate directly into throughput gains without adding more GPUs. For everyday users, the impact is more indirect: faster, cheaper AI services.

Nvidia's 47th filing we've tracked in AI chip wars since July adds to a run that includes GPU handling network tasks and one cutting camera power use.

Source: US 2026/0300779 A1 at the USPTO (PDF). Full patent text and figures are in the official publication.

Get our take in your Top Stories

Liked this breakdown? Add Patentlyze as a preferred source on Google, and our plain-English take shows up more often in your Top Stories the next time Nvidia patent news breaks.

There are more where this came from

We read the new patents from Big Tech, carmakers and more and send you the ones worth knowing. Plain English, free, every week.

The drawings · 48 sheets

48 drawing sheets from US 2026/0300779 A1 · click any drawing to enlarge

Processing system with DPU and GPU., drawing sheet 1 from US 2026/0300779 A1
Sheet 1 Processing system with DPU and GPU.
Input data packets and memory buffers., drawing sheet 2 from US 2026/0300779 A1
Sheet 2 Input data packets and memory buffers.
Inference system with DPU and GPU., drawing sheet 3 from US 2026/0300779 A1
Sheet 3 Inference system with DPU and GPU.
Flowchart of data processing steps., drawing sheet 4 from US 2026/0300779 A1
Sheet 4 Flowchart of data processing steps.
Inference pipeline for input data., drawing sheet 5 from US 2026/0300779 A1
Sheet 5 Inference pipeline for input data.
Network of devices and a server., drawing sheet 6 from US 2026/0300779 A1
Sheet 6 Network of devices and a server.
Data center architecture layers., drawing sheet 7 from US 2026/0300779 A1
Sheet 7 Data center architecture layers.
Network of servers and a remote network., drawing sheet 8 from US 2026/0300779 A1
Sheet 8 Network of servers and a remote network.
Show the other 40 sheets
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 9
Sheet 9
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 10
Sheet 10
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 11
Sheet 11
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 12
Sheet 12
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 13
Sheet 13
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 14
Sheet 14
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 15
Sheet 15
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 16
Sheet 16
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 17
Sheet 17
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 18
Sheet 18
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 19
Sheet 19
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 20
Sheet 20
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 21
Sheet 21
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 22
Sheet 22
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 23
Sheet 23
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 24
Sheet 24
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 25
Sheet 25
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 26
Sheet 26
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 27
Sheet 27
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 28
Sheet 28
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 29
Sheet 29
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 30
Sheet 30
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 31
Sheet 31
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 32
Sheet 32
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 33
Sheet 33
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 34
Sheet 34
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 35
Sheet 35
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 36
Sheet 36
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 37
Sheet 37
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 38
Sheet 38
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 39
Sheet 39
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 40
Sheet 40
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 41
Sheet 41
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 42
Sheet 42
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 43
Sheet 43
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 44
Sheet 44
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 45
Sheet 45
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 46
Sheet 46
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 47
Sheet 47
Nvidia Patents Technology to Format AI Data Inside the Internet Connector Before the GPU Processes It patent drawing, sheet 48
Sheet 48
Patent filing page

Watchlists this filing belongs to

All watchlists →
Reader comments

No comments yet

Start the discussion

Real name or a handle, either is fine. Comments are read by a person before they appear, so allow a little time. Keep it about the filing.