Linux local privilege escalation, AI security and networking updates - Week of May 11, 2026
Companies mentioned
Overview of Recent Activity
Enterprise updates during the reporting period included new Linux local privilege escalation disclosures affecting setuid exploitation paths, multiple security and cloud governance certifications, and vendor updates spanning AI security controls, agentic runtimes, and AI networking protocol deployments. Organizations also issued partnership plans and product releases tied to AI infrastructure and governed AI cloud offerings.
Key Themes and Developments
Technology Releases & Product Enhancements
Fortinet introduced FortiGate 3500G and FortiGate 400G additions to its FortiGate G series, running on FortiOS and using NP7 and SP5 processors, with features including hardware-enforced integrity protections and “native MCP” and agent-to-agent traffic inspection for deeper control over AI data flows, according to the release.
Solo.io added support for NVIDIA’s NemoClaw agent model inside kagent, which it described as a Kubernetes-native agentic runtime with shared runtime controls, observability, and governance for always-on assistants. The company also described NemoClaw as built on OpenClaw, and said the integration extends that pattern into a fleet-scale production runtime.
Partnerships & Ecosystem Engagement
NVIDIA and IREN outlined a partnership for deploying NVIDIA DSX-aligned AI infrastructure, including up to 5 gigawatts over time across IREN’s global data center pipeline. The release also said future deployments were expected to center on IREN’s Sweetwater campus, and described a financial component tied to a five-year right to purchase shares.
Rackspace Technology and AMD signed an MOU for a governed Enterprise AI Cloud for regulated and sovereign workloads, integrating AMD Instinct GPUs and AMD EPYC CPUs into a fully managed stack. The framework described security, governance, and accountability requirements, with Rackspace owning the stack from silicon to outcomes.
Infrastructure, Platform, or Deployment Updates
OpenAI introduced Multipath Reliable Connection (MRC), described as a networking protocol aimed at limiting congestion and failures in large AI clusters and supporting more predictable training across very large XPU systems. The report said MRC extended RoCE and added mechanisms including “multipath packet spraying” and SRv6 source routing, and that OpenAI released the specification through the Open Compute Project.
Datadog reported it achieved FedRAMP High certification for cloud operations, describing it as the top security baseline for protecting controlled unclassified information in cloud environments. The release stated the certification reflects the company’s ability to meet federal cloud security and compliance requirements used for sensitive information.
Full Update Index
VU#260001: Linux kernel contains local privilege escalation vulnerability (Copy Fail); DirtyFrag details two Linux kernel vulnerabilities enabling local root; 650 Group outlines $200B forecast for AI networking market; Netskope outlines capabilities to reduce AI security review delays; NVIDIA and IREN outline DSX AI infrastructure partnership; AMD to Host Annual Meeting of Stockholders; Snyk Integrates Anthropic Claude in Snyk AI Security Platform; Solo.io Adds NemoClaw Support to kagent; ASUS ExpertBook Ultra (B9406CAA) Becomes Available in the US; Rackspace Technology and AMD Sign Memorandum of Understanding to Establish New Category of Governed Enterprise AI Infrastructure; SONiC on software-defined networking for AI-ready foundations; OpenAI’s MRC Initiative Reinforces Ethernet’s Expanding Role in AI Back-end Networks; Datadog Achieves FedRAMP High Certification for Cloud Operations; Wireless Infrastructure Association reelected WIA Board of Directors and Officers; Owl Cyber Defense Honors Carahsoft With 2025 North America Federal Distributor of the Year Award; Fortinet Expands FortiGate G Series to Secure AI from the Data Center to Modern Enterprise Edges; Varonis Systems names Partners in Excellence award winners; Coder Sets a New Standard for AI Coding with Self-Hosted, AI Model Agnostic Coder Agents; Gigamon Reports AI Involved in 83% of Security Breaches; SONiC, automation, and AI Ops for AI traffic patterns; Dell Technologies Swan Principle details hidden AI readiness work; Aviz Networks details ONES demos for open networking at CCISDA 2026; Gartner projects agentic AI for IaC generation and drift remediation; IEEE Honors Ambarella CEO Dr. Fermi Wang With Inaugural 2026 Award.
- VU#260001: Linux kernel contains local privilege escalation vulnerability (Copy Fail)
A local privilege escalation flaw in Linux kernel versions 4.17+ can let an unprivileged user gain root access. The issue, CVE-2026-31431 (“Copy Fail”), involves algif_aead/AF_ALG page-cache writes and targets in-memory setuid binaries. - NVIDIA names Suzanne Nora Johnson to its board of directors
NVIDIA named Suzanne Nora Johnson to its board effective July 13, 2026, and expects her to join the Audit Committee. - DirtyFrag details two Linux kernel vulnerabilities enabling local root
DirtyFrag is a Linux local privilege escalation using CVE-2026-43284 and CVE-2026-43500, with module-blocking mitigation recommended. - 650 Group outlines $200B forecast for AI networking market
AI networking may top $200B by decade’s end, with software estimated at $40–50B, driven by Ethernet fabrics, observability, automation, and open networking. - Netskope outlines capabilities to reduce AI security review delays
The blog describes how AI pilots get stuck in security reviews and outlines Netskope tools for risk scoring, semantic DLP, agent controls, and lower latency. - NVIDIA and IREN outline DSX AI infrastructure partnership
NVIDIA and IREN plan DSX AI factory deployments, including up to 5 gigawatts across IREN’s pipeline and a Sweetwater focus. - AMD to Host Annual Meeting of Stockholders
AMD will hold its 2026 Annual Meeting of Stockholders on May 13, 2026 at 9 a.m. PT as a virtual meeting. The notice includes links for live access and replay, and points to AMD’s 2026 Proxy Statement and Annual Report on Form 10-K for matters to be voted on. - Snyk Integrates Anthropic Claude in Snyk AI Security Platform
Snyk integrated Anthropic’s Claude into its AI Security Platform to automate vulnerability discovery, prioritization, and fixes. - Solo.io Adds NemoClaw Support to kagent
Solo.io added support for NVIDIA’s NemoClaw inside kagent, a Kubernetes-native agentic runtime with security, observability, and governance controls. - ASUS ExpertBook Ultra (B9406CAA) Becomes Available in the US
ASUS made the ExpertBook Ultra (B9406CAA) available for U.S. purchase at $3,599.99, with NPU up to 50 TOPS and NIST-aligned security. - Rackspace Technology and AMD Sign Memorandum of Understanding to Establish New Category of Governed Enterprise AI Infrastructure
Rackspace Technology and AMD signed an MOU outlining a multiyear partnership to build a governed Enterprise AI Cloud for regulated and sovereign workloads. The proposal integrates AMD Instinct GPUs and EPYC CPUs into a fully managed stack, covering private/hybrid deployment, inference runtime, and managed inference services with defined SLAs. - SONiC on software-defined networking for AI-ready foundations
The blog says enterprises should prepare networking foundations for AI by using software-defined platforms like SONiC before adopting AI tools. - OpenAI’s MRC Initiative Reinforces Ethernet’s Expanding Role in AI Back-end Networks
In her latest insight, Sameh Boujelbene examines how OpenAI’s MRC protocol reinforces Ethernet’s growing role in large-scale AI networking infrastructure. Read more>> The post OpenAI’s MRC Initiative Reinforces Ethernet’s Expanding Role in AI Back-end Networks appeared first on Dell'Oro Group. - Datadog Achieves FedRAMP High Certification for Cloud Operations
Datadog reported it achieved FedRAMP High certification for cloud security and compliance requirements tied to CUI protection. - Wireless Infrastructure Association reelected WIA Board of Directors and Officers
WIA reelected its board of directors and extended officer terms after its quarterly meeting in Fort Lauderdale. - Owl Cyber Defense Honors Carahsoft With 2025 North America Federal Distributor of the Year Award
Carahsoft Technology Corp. announced it received Owl Cyber Defense’s 2025 North America Federal Distributor of the Year award. The announcement cites Carahsoft’s sales and marketing work promoting Owl’s cross-domain and data diode solutions for U.S. public sector customers, including via a 2025 Evolution Summit. Availability is referenced through federal contract vehicles. - Fortinet Expands FortiGate G Series to Secure AI from the Data Center to Modern Enterprise Edges
Fortinet announced FortiGate 3500G and 400G additions to its FortiGate G series, citing ASIC-accelerated firewall performance, hardware-enforced integrity protections, and AI-driven security features. The release positions the platforms for data center scale and enterprise edge deployment, with references to FortiOS 8.0 and Fortinet Security Fabric integrations. - Varonis Systems names Partners in Excellence award winners
Varonis named its 2025 Partners in Excellence award winners by global and regional categories, including CDW and GuidePoint Security. - Coder Sets a New Standard for AI Coding with Self-Hosted, AI Model Agnostic Coder Agents
Coder announced a beta of Coder Agents, a native self-hosted AI coding agent. The release positions the product as running the control plane, orchestration, and execution inside customer infrastructure for air-gapped use, centralized governance, and model-agnostic support across providers and self-hosted models. - Gigamon Reports AI Involved in 83% of Security Breaches
Gigamon’s 2026 Hybrid Cloud Security Survey reports AI in 83% of breaches and cites rising breach rates and visibility gaps. - SONiC, automation, and AI Ops for AI traffic patterns
Podcast Episode 20 says AI workloads create bursty east-west GPU traffic, and discusses SONiC, automation, and AI Ops for faster operations. - Dell Technologies Swan Principle details hidden AI readiness work
The post outlines the Swan Principle and lists five infrastructure areas for AI readiness before visible change. - Aviz Networks details ONES demos for open networking at CCISDA 2026
Aviz Networks highlights open networking and AI observability at CCISDA 2026 to improve visibility, reduce costs, and avoid vendor lock-in. - Gartner projects agentic AI for IaC generation and drift remediation
Gartner’s first AI Assistants for IaC guide projects 90% context-aware use and 70% agentic drift remediation by 2029. - IEEE Honors Ambarella CEO Dr. Fermi Wang With Inaugural 2026 Award
IEEE honored Ambarella CEO Dr. Fermi Wang with the inaugural 2026 Arun N. Netravali Video Analytics, Technology and Systems Award.