SLED AI Adoption Intelligence

https://sled.lighthouseadvisory.consulting/feeds/red-hat.xmlSLED AI Adoption Intelligence · Red Hat2026-11-11T05:16:15.449ZLighthouse Advisoryhttps://sled.lighthouseadvisory.consulting/streams/red-hat/editions/2026-11-10Red Hat · Issue 11 · 2026-11-102026-11-11T05:16:15.449Z2026-11-10T00:00:00.000Z

KubeCon NA week puts Red Hat's agent story in front of buyers, but most agent features are still preview or roadmap. Red Hat's blogs show Agent Sandbox and the ACS MCP server as Technology Preview and the Developer Hub 2.1 MCP actions and AI Catalog as developer preview. Red Hat-sponsored interviews say OpenShell is preview and per-agent identity 'comes next'. Independent Tech Times scrutiny finds no maturity labels in the OpenShift 5 release, Severity 3/4-only preview support, no independent testing, and cluster details sent to the chosen LLM. ACS 5 adds CISA KEV prioritization, and a Critical RHEL AI 3.5.3 advisory lists 60 CVEs. Efficiency claims (8x to 10x faster root-cause finding) have no stated method, and no SLED-measured outcome appeared. 6 sources · 3 cross-source patterns Red Hat's agentic launch features are mostly preview, and support tiers surface only in blogs and independent scrutiny Before a pilot or contract, which specific agent features are GA in our exact versions, and what support terms apply to the ones that are not? MCP is becoming the governed interface for agents across OpenShift, ACS and Developer Hub, with RBAC and token scope as the main controls For each MCP endpoint we enable, where is authorization enforced, what data leaves to which model, and who approves write actions? Efficiency claims for AI operations and coding agents arrive without method or public-sector baselines What baseline will we measure on our own incidents and code changes before accepting any vendor efficiency multiple?https://sled.lighthouseadvisory.consulting/streams/red-hat/editions/2026-11-09Red Hat · Issue 10 · 2026-11-092026-11-10T07:05:21.930Z2026-11-09T00:00:00.000Z

Red Hat used KubeCon NA (2026-11-09) to ship OpenShift 5 GA and announce RHEL 10.3/9.9, moving Lightspeed from assistant to operations agent. Independent coverage shows the agentic features remain human-approved and several are Technology or developer preview; OKD docs show AI update assessments send cluster state to the configured LLM. SDxCentral relays Gartner-noted pricing and complexity concerns. Two newly updated OpenShift AI training-component CVEs (Moderate, Low) remain Under investigation. No measured SLED outcomes appeared in-window. 7 sources · 4 cross-source patterns Red Hat AI operations features keep humans approving changes, but support status varies by feature Which Red Hat AI operations features are GA in our exact product version, and what approval workflow governs agent-proposed changes? The LLM endpoint chosen for platform AI determines where sensitive operational data goes Do we have an approved, ideally local, model endpoint and a data-classification decision for cluster and host telemetry before enabling AI operations? OpenShift AI training components show multi-tenant control gaps that need compensating RBAC and admission policy Are job-creation and update permissions on our shared OpenShift AI training clusters aligned with quota and secret access, with admission policy as a backstop? Vendor simple-upgrade messaging contrasts with analyst-reported complexity and pricing concerns What will an OpenShift 5 upgrade and AI-feature adoption cost us in staff time and subscription price over the contract term?https://sled.lighthouseadvisory.consulting/streams/red-hat/editions/2026-11-07Red Hat · Issue 09 · 2026-11-072026-11-08T05:04:20.668Z2026-11-07T00:00:00.000Z

A small, constraint-heavy Red Hat AI edition. Independent TechTarget coverage (5 Nov) of the Red Hat Government Symposium reports a hybrid model strategy: open-weight models on government-controlled infrastructure for routine work, frontier models for selected analysis, with speaker claims (VA OIG: 85% of about 300 AI processes on open-weight models) and an explicit caveat that hosting costs can offset savings. Red Hat's 2 Nov OpenShift AI 3.5 benchmark reports tiered KV-cache offloading gains of up to 5.8x throughput for one long-context model, with smaller or regressing results elsewhere. Red Hat's 3.5 known-issues documentation shows that shared Models-as-a-Service gateways can allow route hijacking by default settings and that rate limiting and llm-d scheduling can fail silently. No new state, local, K-12 or higher-education Red Hat AI deployment evidence was found for 6-7 Nov; all items require local validation. 3 sources · 2 cross-source patterns OpenShift AI 3.5 inference gains and controls must be verified on the customer's stack Which measured tests (trace replay, cache-hit rate, 429 checks, route-attachment checks) will prove that the configured inference optimizations and gateway controls are actually in effect? Self-hosted open-weight economics depend on platform engineering, not token price alone What fully loaded cost per request, including GPU, storage tiers and staff, does the agency achieve on its own routine workloads versus negotiated commercial pricing?https://sled.lighthouseadvisory.consulting/streams/red-hat/editions/2026-11-06Red Hat · Issue 08 · 2026-11-062026-11-07T11:25:39.642Z2026-11-06T00:00:00.000Z

A Data Science Pipelines isolation and control-coverage day for Red Hat AI. Red Hat’s Nov. 5 CVE-2026-88984 (Important) and CVE-2026-88985 (Moderate) map shared-pipeline RBAC bypass and cross-namespace list disclosure to OpenShift AI odh-ml-pipelines images still Under investigation. Independently, upstreambeat covers Open Data Hub 3.6’s modular llm-d/OGX/AI-gateway packaging, while Checkable AI’s 4 Nov OpenShift AI check still marks tamper-proof AI audit and MCP provenance Not addressed. IBM Think’s 6 Nov Lightwell piece centers Red Hat in AI-assisted open-source backport remediation—supply-chain capacity, not a substitute for waiting on RHOAI errata. 5 sources · 2 cross-source patterns Multi-tenant OpenShift AI Data Science Pipelines still fails basic shared-object isolation Before onboarding another agency or research group onto a shared OpenShift AI DSP, have you proven that users without edit roles cannot mutate shared pipelines or read other namespaces’ pipeline records? Upstream modular AI serving expands while independent control checks still show audit and MCP gaps For each new llm-d/OGX/MCP module you plan to enable, which Checkable-style control gaps already have a named compensating owner?https://sled.lighthouseadvisory.consulting/streams/red-hat/editions/2026-11-05Red Hat · Issue 07 · 2026-11-052026-11-06T11:56:04.945Z2026-11-05T00:00:00.000Z

A controls-and-runtime day for Red Hat AI. vLLM's Nov. 4 day-0 support for Moonshot Kimi K3.1 Pro brings measured sparse-attention and AgentX lab gains across NVIDIA, AMD, TPU and Zhenwu—while its quick start enables --trust-remote-code. Red Hat's CVE-2026-114053 record, still listing residual Affected fast-channel, RHEL AI and OpenShift AI packages after the Nov. 3 Inference fixes, states remote code stays disabled by default on Red Hat AI. Separately, IBM Cloud documented Virtual Private Endpoints for Red Hat AI Inference (Nov. 2), and IT Brief's Nov. 5 coverage framed Red Hat–NVIDIA AI Factory around OpenShell agent sandboxes, governed vLLM/llm-d inference and Vera Rubin day-zero plans. No new U.S. SLED deployment evaluation surfaced. 4 sources · 2 cross-source patterns Day-0 model quick starts collide with Red Hat's trust_remote_code default-deny Before any day-0 multimodal model lands on Red Hat AI Inference or OpenShift AI, is trust_remote_code explicitly denied in production ServingRuntimes and covered by a written exception process? Put agent and inference controls outside the model path For each agentic or managed-inference pilot, which controls sit outside the model—sandbox, gateway entitlements and private networking—and who operates them?https://sled.lighthouseadvisory.consulting/streams/red-hat/editions/2026-11-04Red Hat · Issue 06 · 2026-11-042026-11-05T05:43:30.782Z2026-11-04T00:00:00.000Z

A patch-and-provenance day for Red Hat AI. Red Hat shipped Critical Red Hat AI Inference 3.5.3 builds on Nov. 3 covering 60 CVEs, including vLLM code-execution, SSRF and denial-of-service flaws; Red Hat's own data still lists the 'fast' channel images as Affected. RHEL AI also added a packaged GLM-5.3-Flash model from Z.ai, formerly Zhipu AI, whose entities have been on the Commerce Entity List since January 2025, so public buyers need sizing and provenance review. No SLED deployment news or independent evaluation surfaced.

A security-led Tuesday for Red Hat AI. Red Hat pushed 18 OpenShift AI base-image advisories on Nov. 3 that carry the fix for the Brotli flaw CISA lists as known-exploited, extending the Nov. 2 TrustyAI fix to CPU, CUDA, ROCm, Gaudi, Neuron, Spyre and TPU images; one CUDA model-serving image still shows Affected, and sites must rebuild derived images. A Red Hat blog maps post-quantum cryptography onto Red Hat AI Enterprise, useful for NIST-aligned planning but mostly roadmap. No SLED deployment news or independent evaluation surfaced; OpenShift AI 3.6 GA was still not observed.

A thin Monday for Red Hat AI news, led by security operations. Red Hat issued a second OpenShift AI 3.4.6 advisory (RHSA-2026:93550, Nov. 2) fixing the Brotli decoder flaw that CISA lists as known-exploited (federal due date Nov. 10), while its own data still shows one GPU model-serving image as Affected. TechTarget's independent recap of the Red Hat Government Symposium frames AI adoption around legacy systems, overlapping oversight requirements and workforce gaps, but its evidence is federal and quote-driven. OpenShift AI 3.6 GA was still not observed.

A quiet weekend for Red Hat AI news, so this edition covers operational status rather than announcements. Red Hat's own records show that two Important vLLM flaws fixed upstream in vLLM 0.32.0 remain listed as Affected, and five lower-severity ones as Fix deferred, across OpenShift AI, RHEL AI and AI Inference Server. Red Hat says text-only, gateway-protected deployments carry little exposure, and recommends network and image mitigations. A Red Hat field team posted a setup guide for an unreleased operator that adds token metering, budgets and GPU reservations to OpenShift AI's model service; by its own account the guide has not been tested end to end. An independent analysis of llm-d 1.0, the serving layer Red Hat ships, finds useful tenant-priority controls but unfinished request failover and cost-saving claims that rest on narrow assumptions. No new US SLED deployment was found. 3 sources · 1 cross-source patterns Shared Red Hat inference platforms are being governed at the gateway, and much of that control is still the customer's job For a shared inference platform, who sets and tests the gateway rules for authentication, tenant budgets and priority under load, and what evidence shows they work?https://sled.lighthouseadvisory.consulting/streams/red-hat/editions/2026-10-30Red Hat · Issue 02 · 2026-10-302026-10-31T04:40:53.864Z2026-10-30T00:00:00.000Z

Red Hat's AI messaging this week centered on agents acting through governed layers, and this edition's evidence is almost entirely Red Hat-authored. An architect argues that AI recommendations only pay off when they trigger reviewed Ansible automation, and Red Hat's documentation confirms the Ansible MCP server has been generally available since June, though same-day coverage called it a preview. The documentation also warns that read-write mode risks hallucinated commands and that tool results go to external LLM providers unfiltered. A proof of concept tests role-based agent data access from written policy. A Red Hat-commissioned EMEA survey finds governance confidence outrunning visibility. Partner Atos reports that small models tuned on OpenShift AI matched a frontier teacher for a German public-sector customer, with undisclosed evaluation details. Red Hat also joined an NSF-backed UC Berkeley agent-efficiency project. No new US SLED deployment or independent evaluation was found. 6 sources · 2 cross-source patterns Red Hat's agent story is shifting from what agents can do to which governed layer they act through For each proposed agent, what governed layer does it act through, what token or role scopes it, and what negative test proves it cannot exceed that scope? Confidence in AI governance is running ahead of verification What evidence, beyond policy documents and vendor assessments at onboarding, shows that AI data access, dependencies and providers are being verified continuously?https://sled.lighthouseadvisory.consulting/streams/red-hat/editions/2026-10-29Red Hat · Issue 01 · 2026-10-292026-10-30T04:54:01.797Z2026-10-29T00:00:00.000Z

First red-hat edition. Red Hat's AI story this month centers on agent safety and sovereign, self-hosted AI, while the independent evidence centers on the operating work those claims require. Red Hat and NVIDIA announced the OpenShell agent-sandbox reference architecture, and an independent source audit found the design sound but the defaults permissive. A Critical OpenShift AI 3.5.2 errata and an upstream vLLM code-execution advisory show the patching load of self-hosted AI. A preprint found unguided AI-generated Ansible consistently insecure. On the deployment side, UNC's on-premises health assistant and a German classified AI cloud remain pre-production, and New England's research cloud paused its OpenShift AI services during a platform rebuild. 8 sources · 3 cross-source patterns Agent safety is being sold as a platform layer, but its defaults need hardening Before hosting agents on Red Hat AI, who will own sandbox policy, enforce non-default settings, and prove egress and credential controls with negative tests? Self-managed Red Hat AI makes patching and upstream tracking a customer duty Does the agency have a patch SLA, an inventory of inference endpoints and an owner who maps upstream advisories to deployed Red Hat builds? Institution-controlled AI depends on approvals, staffing and continuity, not just the platform Who will fund, staff, accredit and sustain the self-hosted AI service after the pilot, and how will users move if the platform changes?