The Circuitry
THE CIRCUITRYYour one-stop source for all tech news
HOMETODAYNEWSFEEDEVENTS
BOOKMARKS
SOONSTARTING SOONMeta Q3 2026 Earnings CallIN 7HOpen coverage →
RSS
© 2026 The Circuitry
About UsSourcesContactCorrectionsPrivacy
  • Today
  • Feed
  • Events
  • Saved
Scroll for more
Verification
VERIFIEDConfidence: HIGH
Source identified
Claims cross-referenced
No discrepancies found
Sourcing
1source

via NVIDIA Newsroom

NVIDIA Newsroom · track record
2Stories
100%Verified
230d
All sources →
Markets
NVDA···

Live quote · not investment advice

From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
Home/Tech/NVIDIA Launches OpenShell Reference Architecture to Contain and Police Enterprise AI Agents
VERIFIEDBy Xavier Rivera·Wednesday, October 21, 2026 at 6:19 AM EDT·2 min read

NVIDIA Launches OpenShell Reference Architecture to Contain and Police Enterprise AI Agents

NVIDIA's new OpenShell Reference Architecture is a full hardware-and-software stack for running AI agents inside enforced boundaries, with BlueField-4 chips able to quarantine a misbehaving agent in milliseconds. It is available now, and Anthropic, OpenAI, Microsoft, Google Cloud, AWS, CrowdStrike and Palo Alto Networks are among the companies working with it.

Source:NVIDIA Newsroom
Post
NVIDIA Launches OpenShell Reference Architecture to Contain and Police Enterprise AI Agents
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
NVIDIA on Wednesday announced the NVIDIA OpenShell Reference Architecture at GTC Berlin. It is a full-stack design for building, containing, observing and enforcing safe AI agents on enterprise infrastructure and edge systems.

It extends NVIDIA's open source OpenShell secure runtime software into what the company calls industrial-scale infrastructure for developing and running agents.

Three layers: think, act, govern
The architecture has three layers:
Vera Rubin runs model inference, for both closed models and open models built with NVIDIA Nemotron. For agentic workloads, NVIDIA says Vera Rubin NVL72 systems deliver 30x higher throughput per megawatt.
Vera CPUs run the agent harnesses and tool use, along with OpenShell, which sets a secure runtime boundary for autonomous agents. NVIDIA claims Vera CPUs deliver 2x the performance of x86 processors with greater energy efficiency.
BlueField-4 DPUs run NVIDIA Sentry, a security framework built on NVIDIA DOCA. It monitors agent behavior continuously and out-of-band, with in-silicon policy enforcement. If an agent tries to move outside its software boundary, Sentry quarantines and stops it in milliseconds. NVIDIA says Sentry also boosts agent capacity by up to 20%.
From The CircuitryThe Feed — live briefs across tech, all day.See what’s happening →
The design extends to the edge with NVIDIA RTX Spark, DGX Spark and DGX Station systems. NVIDIA lists retail, healthcare, manufacturing, robotics, software development and telecommunications as target industries.

"AI capability and safety must advance together," said Jensen Huang, founder and CEO of NVIDIA. "As enterprises prepare to deploy billions of AI agents, safety and control must be built into the computing architecture. With the OpenShell Reference Architecture, agents can think on Rubin, act on Vera and be governed on BlueField."

Ecosystem
Operating systems include Linux and Kubernetes infrastructure from Canonical, SUSE and Red Hat, and Microsoft Windows via Microsoft Execution Containers (MXC). Clouds include AWS, Google Cloud and Microsoft Azure. Enterprises can run agents from Anthropic, Cohere, Google, Meta, OpenAI, Perplexity, Reflection and SpaceX Super Intelligence on OpenShell infrastructure. Security vendors including 1Password, Check Point, CrowdStrike, Cisco, F5, Fortinet, Microsoft, Okta, Palo Alto Networks, TrendAI and Zscaler are integrating support into their agent security products. Europe's BBVA and Santander are exploring the architecture with NVIDIA. System makers including Dell Technologies, Cisco, HPE, Lenovo and Supermicro, and cloud providers including CoreWeave, Crusoe, Firebird, Firmus, GMI Cloud, IREN, Lambda, Nebius, Nscale, Oracle Cloud Infrastructure and Vultr, are working with NVIDIA on OpenShell-ready infrastructure.

The NVIDIA OpenShell Reference Architecture is now available.
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
Why this mattersAI · ~100 words

Tap a lens to see what this story means for you.

Morning Brief

Liked this? The Brief brings you the whole day in tech, verified, every morning.

Two minutes, free forever. What's in The Brief →

Reader-supported
DonateBuy me a coffee →Follow@thecircuitry_ →Follow@thecircuitry.to →
HELP US IMPROVE
From The Circuitry

See what’s happening right now

The Feed runs all day — short, verified briefs the moment they break.

Open the Feed →
From The Circuitry

Follow @thecircuitry_

Every story we publish, as it happens. No noise between.

Follow on X ↗On Bluesky ↗

Reader-supported

The Circuitry is a passion project I've always wanted to build, and I love the work behind it.

Running it costs real money. APIs, hosting, time. To keep improving the site and growing this into something useful for everyone, those costs have to be covered.

Any contribution is appreciated. If not, no pressure. Thanks for reading.

Buy me a coffee
More fromNVIDIA Newsroom
  • NVIDIA Introduces DSX POD, Factory-Built Vera Rubin AI Factories That Deploy in as Little as 14 Days

    Tech · 2h
More inTech
  • T-Mobile Completes First AI-RAN Call on a Commercial Network With NVIDIA and Nokia

    Tech · 1h
  • NVIDIA and UW Institute for Protein Design Release RFD4-Proteina Open Model

    Tech · 1h
  • Bolt, Mercedes-Benz, VW and Einride Build Autonomous Vehicles on NVIDIA Hyperion

    Tech · 1h
SupportThe Work

The Circuitry is reader-supported. If you find the daily brief useful, you can buy me a coffee to keep it going.

Buy a coffee →
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →

MORE IN THIS BEAT

All Tech →
  • Tech·Oct 21

    Mistral says it is working with NVIDIA on AI for industrial engineering

    Mistral said on Wednesday it is collaborating with NVIDIA on AI-native engineering workflows, including faster simulation, public benchmarks and open datasets for physics AI, and evaluations for agentic AI in CAE and EDA.

  • Tech·Oct 21

    Oracle patches 27 SOA Suite flaws, three rated CVSS 10.0

    Oracle's October 2026 Critical Patch Update fixes 27 SOA Suite vulnerabilities, three of them rated CVSS 10.0. The three maximum-severity flaws need no authentication, and Oracle says they could allow takeover of affected versions 12.2.1.4.0 and 14.1.2.0.0.

  • Tech·Oct 21

    Oracle Access Manager hit by CVSS 10 critical flaw

    Oracle disclosed CVE-2026-87337, a CVSS 10.0 flaw in Access Manager versions 12.2.1.4.0 and 14.1.2.1.0 that allows unauthenticated remote attackers to compromise data confidentiality and integrity. The maximum-severity issue carries scope change and affects the Authentication Engine component of Oracle Fusion Middleware.

  • Tech·Oct 21

    NVIDIA GTC Berlin 2026 features robotics and infrastructure updates

    NVIDIA announced Isaac platform expansion with Agile ONE SN and DSX POD at GTC Berlin 2026. T-Mobile completed the first AI-RAN call on a commercial network using NVIDIA technology.

  • Tech·Oct 21

    NVIDIA Expands Isaac Robotics Platform With Agile ONE SN Reference Humanoid

    NVIDIA is expanding its Isaac open robotics platform with Agile ONE SN, a reference humanoid from Agile Robots built on Isaac GR00T and Jetson Thor, plus a new Isaac OS runtime and Isaac Lab 3.0. Agile ONE SN is scheduled to be available in early 2027.