Smart Network Interface Offload Patents: Top Companies & Trends 2026
- 89.5% of all 973 records sit with just five assignees, leaving a long tail of single- or double-digit filers behind them.
- Filings peaked in 2022 at 251 then fell to 99 by 2024, a -37% swing over that three-year span (2025-2026 figures are still filling in due to publication lag).
- G06F and H04L dominate at 56.1% and 41.2% of records respectively, while wireless (H04W) and recognition classes (G06K, G06V) stay under 2% each.
Filing growth compares 2021 (157 records) with 2024 (99) — a three-year span. 2024 is the most recent year we treat as complete: publication lags filing by roughly 18 months, so 2025 onwards are still filling in and any growth rate that ends there would understate the field. Top-5 share is the combined record count of the five largest assignees divided by all 973 records in scope (CR5), not by the ranked leaders only.
What this landscape covers
Smart network interface offload spans patent activity around programmable NICs, transport offload engines, match-action packet pipelines, and the host-interface and tenant-isolation mechanisms that let a network card do work a CPU used to do. This dataset covers 973 published records filed or published between 2015 and mid-2026, drawn from a search string built around offload, programmable pipeline and telemetry/isolation terms appearing together in title, abstract, claims or description.
The records are grouped into patent families for the assignee ranking, so continuation filings and multi-jurisdiction copies of the same invention are counted once rather than inflating any single company's total. Because publication lags filing by roughly 18 months, the 2025 and 2026 counts in the trend chart are necessarily incomplete and should not be read as a slowdown.
Filing trend and technology composition
Two views of the same 973 records: how filing volume moved year over year, and which IPC subclasses the underlying inventions actually sit in.
Filing trend, 2017-2026
Annual filings rose from 12 in 2017 to a peak of 251 in 2022, then declined to 99 by 2024, a documented -37% change over that three-year span. The 2025 and 2026 figures shown are partial because publication trails filing by about 18 months; they should not be read as evidence the field is cooling further.
IPC subclass composition
G06F (electric digital data processing) appears on 56.1% of records and H04L (digital information transmission) on 41.2%, reflecting the core split between on-NIC compute and network transport. G06T (image data processing) reaches 26.3%, driven by records that pair packet offload with media or vision pipelines. Since a single record can carry several IPC codes, these shares add up to well over 100% of the 973-record total, which is expected and not an error.
Shares are the percentage of the 973 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Smart Network Interface Offload with Eureka
This page is one run against one query. Ask Eureka your own question about smart network interface offload and every answer comes back with the patent numbers behind it.
Try EurekaMost-cited records and a representative filing
US20250068457A1 — Computer vision pipeline management in programmable network interface device
The application describes an apparatus with a host interface, a network interface, and programmable circuitry that decides how to split computer vision processing between the programmable NIC and the host device, then deploys the relevant portions accordingly.Filed by Intel, published 2025-02-27 — illustrates how offload claims are extending from pure packet/transport processing into on-NIC media and vision workload placement.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US20210144517A1 | Multi-entity resource, security, and service management in edge computing deployments | 840 |
| 2 | US20130044641A1 | Federating interconnection switching element network to two or more levels | 376 |
| 3 | US20130044764A1 | Generating flows for managed interconnection switches | 242 |
| 4 | US20130058335A1 | Packet processing for logical datapath sets | 219 |
| 5 | US20210117360A1 | Network and edge acceleration tile (NEXT) architecture | 211 |
| 6 | US20200322287A1 | Switch-managed resource allocation and software execution | 178 |
| 7 | US20170180273A1 | Accelerated network packet processing | 143 |
| 8 | US20150016266A1 | Traffic Management with Ingress Control | 126 |
| 9 | US20130044751A1 | Packet processing in managed interconnection switching elements | 122 |
| 10 | US20050226238A1 | Hardware-based multi-threading for packet processing | 113 |
Citation counts inside a searched corpus favour older filings that have had more time to accumulate citations; treat them as a signal of influence within this dataset, not as a ranking of current technical importance.
Each row carries its publication number; clicking a row searches Eureka by that number.
Put your own technology through the same analysis
Eureka on the web
When you want the answer in the next five minutes.
The agent works the prompt against patents and technical literature, citing every source.
Run your analysis now →MCP server & REST API
When it has to run inside your own pipeline.
Patent search, landscape analysis and assignee resolution as MCP tools. Drop them into any agent framework, or call REST directly.
Browse MCP servers →What the numbers mean for a filing decision
Three patterns stand out once the records are grouped by assignee, class and year — each has a direct bearing on where new claims are likely to survive prosecution.
Claim space at the top is effectively closed
With five assignees holding 871 of the 973 records in scope, the core offload, transport and pipeline claim territory is heavily occupied. New entrants filing head-on claims in this space face dense prior art from the leader alone, whose total dwarfs every other ranked assignee.
Volume has pulled back from its 2022 peak
Filings rose sharply through the late 2010s, peaked at 251 in 2022, and had fallen to 99 by 2024 — a documented -37% change across that three-year span. That is a real pullback in completed filing years, not an artefact of publication lag, since 2024 is the most recent year that can be treated as complete.
Compute and transport classes dominate
Over half of records touch general digital-processing claims (G06F) and over two in five touch transmission claims (H04L), with image-processing crossover (G06T) present in roughly a quarter. Wireless, recognition and multiplex classes each sit under 2%, marking them as comparatively open ground.
A small inventor cluster anchors the leader's portfolio
The strongest co-filing relationship in the dataset links two named inventors across 17 shared records, with two further pairs at 15 each — all clustered around the same core group. That concentration suggests a compact original inventing team behind a large share of the leading assignee's filings.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to smart network interface offload, with the prior art for and against each one.
Who is filing, and where activity has cooled
The ranking covers 78 companies in total — not a top-50 or top-100 cut, but the complete set the data endpoint returns. Recent-year momentum figures show where filing has slowed even among the leaders.
One assignee dominates by volume
The leading assignee's 693 records alone account for the majority of all 973 in scope, far ahead of fifth place at 27 and tenth place at 7. That gap between first and fifth is itself evidence of how top-heavy this field is.
Even the leader's recent filing has slowed sharply
The top-ranked assignee filed just 3 records in the latest year, down 97% year-on-year, and several other ranked assignees show zero filings in the same period. Given the 18-month publication lag, some of this reflects filings still in the pipeline rather than a definitive stop.
A long tail sits well below the leaders
Beyond the top 10, which together hold 95.1% of all 973 records, the remaining assignees each hold small single-digit counts. That distribution points to a field where challenging the leaders directly is costly, but niche filings around specific mechanisms still find room.
| Assignee | Recent year | YoY |
|---|---|---|
| Intel Corporation | 3 | -97% |
| Mellanox Technologies Ltd. | 1 | -75% |
| Cisco Systems, Inc. | 0 | — |
| Microsoft Technology Licensing, LLC | 0 | -100% |
| Nicira, Inc. | 0 | — |
| KOPONEN TEEMU | 0 | — |
| Xilinx, Inc. | 0 | — |
| CASADO MARTIN | 0 | — |
Where to take this next
The dataset points to a field with closed core claim territory and a few specific open branches. The next steps depend on whether the goal is freedom-to-operate, acquisition targeting, or drafting around the leaders.
Map freedom-to-operate against the top 5
Given that 89.5% of records sit with five assignees, any new filing in core offload or transport-pipeline claims should be checked against that group's portfolios first, not the full 78-company ranking.
Run an FTO scan in EurekaWatch the under-claimed branches
Wireless-transport integration and on-NIC vision pipelines show comparatively low filing density; drafting claims there faces less prior art but also less validated demand signal.
Explore white space in EurekaTrack the pullback from 2022 peak filing
Volume fell from 251 in 2022 to 99 in 2024. Monitoring whether 2025-2026 filings (once fully published) confirm or reverse that trend will matter for timing new applications.
Set up trend alerts in EurekaCommon questions about this landscape
One assignee holds 693 of the 973 records in scope, a scale far ahead of fifth place at 27 and tenth place at 7. The top 5 assignees combined hold 89.5% of all records, and the top 10 hold 95.1%, so the field is heavily concentrated at the very top rather than spread evenly across the 78 ranked companies. Anyone assessing competitive risk in this space should treat that leading assignee's portfolio as the primary reference point, not the ranking as a whole.
Filings grew steadily through the late 2010s, peaked at 251 in 2022, and fell to 99 by 2024 — a documented -37% change across that span. Because publication lags filing by roughly 18 months, 2024 is the most recent year that can be treated as complete, and the lower counts shown for 2025-2026 are still filling in rather than confirmed. Based on completed years, the field has pulled back meaningfully from its 2022 peak.
The largest share, 56.1% of the 973 records, carries a G06F classification for general digital data processing, and 41.2% carries H04L for digital information transmission, reflecting the split between on-card compute and network transport claims. A notable 26.3% also touches G06T image-processing classes, showing crossover into media and vision workloads on programmable NICs. Wireless (H04W), recognition (G06K) and multiplex (H04J) classes each sit under 2%, marking comparatively lighter-claimed territory.
The technology composition data shows wireless-transport integration, on-NIC vision or recognition pipelines, and multiplex-channel telemetry all sitting under 2% of records, well below the dominant compute and transport classes. That does not guarantee those areas are commercially unproven, but it does mean fewer competing claims to navigate. A first filing there would need to tie the offload mechanism specifically to one of these lighter-claimed functions rather than restating general packet-pipeline or transport-offload language already covered by the leading assignees.
The most-cited records in this dataset, including one with 840 citations, were filed and published earlier in the coverage window, which gives them more time to accumulate citations inside the searched corpus than recent filings could ever have had. That makes citation count a useful signal of historical influence on how the field's foundational concepts were framed, but a weak signal of which claims matter most for a 2026 filing decision. Recent representative filings, such as the 2025 computer-vision-on-NIC application, show where claim drafting is currently heading even though they carry few citations yet.
Research Smart Network Interface Offload in depth with Eureka
Go past this page: query the whole smart network interface offload corpus yourself, in your own scope.
Every answer comes back with patent numbers you can open.
Disclaimer. This page is generated from Patsnap Eureka data drawn from a limited snapshot of global patent and scientific-literature records, and is provided for general information and reference only.
Patent data carries inherent limitations: recent filings (typically the most recent 18–24 months) are under-counted due to standard publication lag; counts may be reported at either a patent-family or a patent-record basis and are not always directly comparable; classification, applicant-name, and citation data may contain errors, duplicates, or omissions; and the underlying search query defines and constrains the scope shown. As a result, the analysis may be incomplete or inaccurate and may not reflect the full technology landscape.
Nothing on this page constitutes an exhaustive prior-art, novelty, freedom-to-operate, or validity search, nor does it constitute legal, financial, investment, or professional advice, and it should not be relied upon as such. Any patent, commercial, or strategic decision should be verified independently and reviewed with qualified patent, legal, and domain professionals. Patsnap makes no warranties, express or implied, as to the accuracy, completeness, or fitness for any particular purpose of the information presented.
Machine translation. Assignee and organisation names originally recorded in Chinese, Japanese or Korean have been rendered into English by an AI translation step so that the tables stay readable. These renderings are best-effort and may not match a company’s registered English name; the original name is what the underlying patent record carries, and it is what any Eureka query launched from this page uses.