Skip to main content
SambaNovaPatent grant

SambaNova Systems was granted a patent for flexible AI data movement.

Source: Autobound Signal API · 70 signals tracked for SambaNova · updated

Get SambaNova's live signals in Claude or ChatGPT, or by API

AskWhat changed at SambaNova this week? Use Autobound signals for sambanova.ai.

What happened

In November 2026 SambaNova Systems patented memory-management technology for AI accelerators and data-processing systems that move information across connected computing hardware.

Source

Abstract

A pattern memory unit (PMU) connected to a switch fabric can implement bypassable watchpoint contexts that send scratchpad data from a scratchpad memory in the PMU to the switch fabric in response to receiving a first trigger. The PMU also implements an input context and an output context. The input context writes scratchpad data to the scratchpad memory in response to receiving input data from the switch fabric. The output context reads the scratchpad data from the scratchpad memory and sends output data to the switch fabric in response to receiving a second trigger. Bypassing the bypassable watchpoint context includes intercepting the first trigger and sending the second trigger to the output context without running the bypassable watchpoint context.

patents.google.com/patent/US12792302Read the full source

Extracted by Autobound

From the Signal API record
Signal
Patent grant

What this signalsA new patent often shows where a company is putting its R&D budget.

Grant date
Nov 10, 2026
Inventors
3
Tech area
software / computing
Patent number
US12792302

The full record

From the Signal API record

People

  • Saubhik MukherjeeInventor
  • Matthew FeldmanInventor
  • Kaushik Bharadwaj ManchellaInventor

Details

Primary CPC class
G06F 7/36 (Electric digital data processing)
USPTO assignee
SambaNova Systems, Inc.
What the invention does
A pattern-memory design that stores incoming data locally and, when triggered, can either inspect it or send it directly onward through the chip network.

Topics and mentions

Tags

  • Intellectual Property
  • Research & Development
  • Artificial Intelligence
  • Engineering

Technologies named

  • AI accelerators
  • pattern memory
  • scratchpad memory
  • data fabrics

Extraction

Detected
Nov 11, 2026
signal_type
patents-company
signal_subtype
patentGrant

Use this data

Get every patent signal for SambaNova and the companies you sell to, in the tools you already use.

  1. Ask Claude about it

    Connect Autobound to Claude, Claude Code or Cursor with MCP. Then ask: “What changed at SambaNova this week?”

  2. Send it to your own tools

    The Signal API returns patent signals for any list of companies as JSON, for your CRM, warehouse or app.

  3. Try it free

    Sign up and spend your free credits on the companies you sell to.

    Start Free1,000 free credits

The API returns more than this page shows

This page shows a preview. The full patents-company record in the Signal API and MCP can also have these 8 fields. Some fields are empty for some signals.

Company

  • linkedin_urlValue in the API
  • industriesValue in the API
  • employee_count_lowValue in the API
  • employee_count_highValue in the API
  • revenueValue in the API
  • descriptionValue in the API

Signal

  • signal_nameValue in the API
  • associationValue in the API
Show the full JSONThe record on this page and the API request

GET /v1/signals/8a39ab20-eaeb-54ce-ba19-c80510e2c4b2 returns this record as JSON. POST /v1/companies/enrich returns every signal for sambanova.ai.

{
  "signal_id": "8a39ab20-eaeb-54ce-ba19-c80510e2c4b2",
  "signal_type": "patents-company",
  "signal_subtype": "patentGrant",
  "detected_at": "2026-11-11T15:52:53+00:00",
  "company": {
    "name": "SambaNova",
    "domain": "sambanova.ai"
  },
  "data": {
    "cpc": [
      {
        "code": "G06F 7/36",
        "label": "Electric digital data processing"
      }
    ],
    "tags": [
      "Intellectual Property",
      "Research & Development",
      "Artificial Intelligence",
      "Engineering"
    ],
    "detail": "In November 2026 SambaNova Systems patented memory-management technology for AI accelerators and data-processing systems that move information across connected computing hardware.",
    "summary": "SambaNova Systems was granted a patent for flexible AI data movement.",
    "event_at": "2026-11-10",
    "inventors": [
      {
        "name": "Saubhik Mukherjee",
        "status": "resolved",
        "is_primary": true,
        "linkedin_url": "https://www.linkedin.com/in/saubhikm"
      },
      {
        "name": "Matthew Feldman",
        "status": "unresolved",
        "is_primary": false
      },
      {
        "name": "Kaushik Bharadwaj Manchella",
        "status": "unresolved",
        "is_primary": false
      }
    ],
    "tech_area": "software / computing",
    "event_kind": "grant_date",
    "grant_date": "2026-11-10",
    "source_url": "https://patents.google.com/patent/US12792302",
    "cpc_primary": {
      "code": "G06F 7/36",
      "label": "Electric digital data processing"
    },
    "patent_title": "Systems and methods for bypassable watchpoint contexts in pattern memory units",
    "patent_number": "US12792302",
    "uspto_assignee": "SambaNova Systems, Inc.",
    "patent_abstract": "A pattern memory unit (PMU) connected to a switch fabric can implement bypassable watchpoint contexts that send scratchpad data from a scratchpad memory in the PMU to the switch fabric in response to receiving a first trigger. The PMU also implements an input context and an output context. The input context writes scratchpad data to the scratchpad memory in response to receiving input data from the switch fabric. The output context reads the scratchpad data from the scratchpad memory and sends output data to the switch fabric in response to receiving a second trigger. Bypassing the bypassable watchpoint context includes intercepting the first trigger and sending the second trigger to the output context without running the bypassable watchpoint context.",
    "precision_class": "structural",
    "invention_explanation": "A pattern-memory design that stores incoming data locally and, when triggered, can either inspect it or send it directly onward through the chip network.",
    "technologies_mentioned": [
      "AI accelerators",
      "pattern memory",
      "scratchpad memory",
      "data fabrics"
    ]
  }
}

Long text fields are shortened on this page.

Looking up one signal by its id is free. Enrich costs 2 credits per signal returned; a call with no results is free.