惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

S
SegmentFault 最新的问题
Google DeepMind News
Google DeepMind News
G
Google Developers Blog
Martin Fowler
Martin Fowler
MongoDB | Blog
MongoDB | Blog
月光博客
月光博客
Jina AI
Jina AI
宝玉的分享
宝玉的分享
人人都是产品经理
人人都是产品经理
D
DataBreaches.Net
V
V2EX
WordPress大学
WordPress大学
T
The Blog of Author Tim Ferriss
Last Week in AI
Last Week in AI
B
Blog
博客园 - 叶小钗
小众软件
小众软件
Stack Overflow Blog
Stack Overflow Blog
P
Proofpoint News Feed
A
About on SuperTechFans
J
Java Code Geeks
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
Y
Y Combinator Blog
Microsoft Security Blog
Microsoft Security Blog

Ubuntu blog

Beyond the 10-year mark: Extending Ubuntu Pro 16.04 LTS security coverage | Ubuntu Android™ development shouldn’t start with a physical device | Ubuntu Bring Zenoh to ROS 2 with snaps | Ubuntu Evolution of the RISC-V ISA. What next after RVA23? | Ubuntu Cut bloat, not features | Ubuntu Ubuntu now certified on Qualcomm Dragonwing™ IQ-8275 | Ubuntu Grace on the currents: Stonking Stingray | Ubuntu How we create a Canonical Academy exam | Ubuntu Surviving the uncharted: when dedicated OpenStack expertise is your best ally in disaster recovery  | Ubuntu Canonical joins the Open Secure AI Alliance | Ubuntu AI harnesses for telco autonomous networks | Ubuntu Arduino® VENTUNO™ Q is available for pre-order with Ubuntu pre-installed | Ubuntu Advantech AOM-2721 is now Ubuntu Certified | Ubuntu Canonical integrates NVIDIA Nemotron 3.5 Lightning with Ubuntu for always-on AI agents | Ubuntu Ubuntu’s virtualization hardware enablement (HWE) stack: a new model for confidential computing enablement | Ubuntu Confidential computing and the new regulatory focus on data in use | Ubuntu A day in the life of an Android developer with Anbox Cloud | Ubuntu Canonical announces the Enterprise Store as part of Ubuntu Pro | Ubuntu Tracing a memory leak bug in PID 1 and contributing an upstream fix: a Linux support story | Ubuntu MAAS installation: bare metal provisioning is easier than ever | Ubuntu Januscape vulnerability CVE-2026-53359 mitigations available | Ubuntu Managing Ubuntu on bare metal at scale | Ubuntu Ubuntu Server: a platform made for enterprise scale | Ubuntu Building an open source chain of trust: new research uncovers key blockers and ways forward | Ubuntu Beyond safety and security: Why automotive open source demands dependability  | Ubuntu DirtyClone Linux kernel local privilege escalation vulnerability fixes available | Ubuntu pedit COW kernel local privilege escalation vulnerability mitigations | Ubuntu Canonical becomes Gold Sponsor of Trifecta Tech Foundation | Ubuntu Challenges designers face in open source (and how to fix them) | Ubuntu Hunting a 16-year-old SQLite bug with TLA+: is dqlite affected? | Ubuntu
Run agentic workloads on Arm and Ubuntu | Ubuntu
Youssef Elto · 2026-05-26 · via Ubuntu blog

The era of prompt-and-response AI is behind us. We are now firmly in the age of agentic AI and the world needs a new class of compute built for this reality. That is why the launch of the Arm® AGI CPU is a pivotal moment. And to harness that raw, bare-metal efficiency at scale, it requires an optimized, resilient, and securely designed operating system.

In the lead-up to Ubuntu Summit 26.04, Canonical and Arm are collaborating to certify the new Arm AGI CPU on Ubuntu 26.04 LTS (Resolute Raccoon).

A shared history: pioneering Arm infrastructure

Canonical didn’t just jump on the Arm bandwagon; we have been pioneering emerging architectures for decades. Our vision, as stated in the historic 2012 Ubuntu Server 12.04 LTS press release, was a “future in which low-energy, hyperscale servers come to dominate for many workloads.” Ubuntu 12.04 LTS was the first enterprise Linux distribution to provide support for Arm-based servers, laying the early groundwork for the hyperscale era. From those early 32-bit development boards, Canonical has treated Arm as a first-class citizen, evolving Ubuntu into the dominant OS for the 64-bit server market.

As hyperscalers transitioned to cloud-native Arm processors in 2018, Ubuntu became the go-to foundation, providing day 0 support for all public cloud Arm instances. Canonical’s commitment goes beyond the operating system itself. Our infrastructure suite – including MAAS, Openstack, MicroCloud, Ceph, LXD, and Canonical Kubernetes – has featured highly optimized, native Arm64 support for years. This has enabled seamless bare-metal provisioning and orchestration, ensuring that when new Arm64 silicon drops, the software ecosystem is already waiting.

The new benchmark: Arm AGI CPU

For 35 years, Arm has licensed its IP to power the world’s computing. The introduction of the Arm AGI CPU expands this platform strategy, giving the ecosystem the ultimate flexibility to build on Arm – whether licensing IP, adopting Arm Neoverse® CSS, or deploying Arm-designed silicon. Arm AGI CPU is a high-performance engine built specifically to address the power and density constraints of modern AI workloads.

Hardware highlights include:

  • Massive compute density: Up to 136 Arm Neoverse V3 cores spread across a dual-chiplet design.
  • Efficiency: Manufactured on TSMC’s 3nm process, delivering maximum throughput at 300W TDP.
  • Sub-100ns Memory Latency: 12 channels of DDR5 memory support, delivering 6GB/s of memory bandwidth per core to eliminate data starvation in real-time AI inference.
  • Rack-scale economics: The 1OU dual-node reference design packs 272 dedicated cores per server unit, completely reshaping data center economics.

“The shift to agentic AI demands a new class of infrastructure where performance, efficiency, and scalability are engineered together from the ground up. By collaborating with Canonical to certify Ubuntu 26.04 LTS on Arm AGI CPU, we are enabling developers and enterprises to deploy next-generation AI workloads with the performance density, power efficiency, and software readiness required for the agentic era.”— Eddie Ramirez, VP of Go-to-Market, Cloud AI Business Unit, Arm

Ubuntu 26.04 LTS on Arm AGI CPU: hyper-efficient AI infrastructure

In high-stakes data centers, silicon is only as powerful as the software running on it. To extract every ounce of performance from the Arm AGI CPU, Ubuntu 26.04 LTS – backed by Canonical’s full-stack infrastructure suite – provides a secure, resilient, and frictionless foundation.

  • Cutting-edge kernel 7.0 with 15-year support: Ubuntu 26.04 LTS ships with the Linux 7.0 kernel, bringing the latest upstream kernel innovations directly to enterprise environments. Backed by up to 15 years of enterprise maintenance with Ubuntu Pro, this release natively unlocks advanced Arm64 features like MPAM and nested virtualization.
  • Arm64 kernel livepatching: In the world of continuous agentic AI, downtime is unacceptable. For the first time, Canonical’s Kernel Livepatch is available for Arm64 servers, allowing operators to apply critical security updates to the kernel without ever rebooting.
  • Memory safety at the core: Ubuntu 26.04 LTS integrates Rust into the system layer (such as sudo-rs and coreutils), aligning perfectly with hardware-enforced security capabilities, like MTE, built into the Neoverse V3 architecture.
  • Out-of-the-box optimization: Ubuntu certification brings rigorous, continuous validation in Canonical labs for the entire infrastructure portfolio to deliver secure, stable, and highly performant day-one deployments on certified hardware.

“The long-standing collaboration between Canonical and Arm has consistently driven  the evolution of enterprise infrastructure. Our joint efforts to certify the Arm AGI CPU on Ubuntu 26.04 LTS ensures a secure, dependable foundation to orchestrate agentic AI in data centers and cloud”— Cindy Goldberg, VP of Cloud and Silicon Partnerships at Canonical

Join Us at Ubuntu Summit 26.04

We invite you to join us at Ubuntu Summit 26.04 to see the future of infrastructure firsthand. Attend Arm’s dedicated workshop on optimizing workload on Arm and Ubuntu through agentic AI and static performance analysis for application tuning.

Register for Ubuntu Summit 26.04