CNCF

Upbound Model Plane: Fleet-Wide AI Inference Orchestration

Upbound Model Plane: Fleet-Wide AI Inference Orchestration

Upbound announced Model Plane at KubeCon, a fleet-wide AI inference control plane built on Crossplane and donated to the CNCF. The project lets enterprises manage AI model deployments across fragmented GPU sources using the same Kubernetes-native operational model they already use for cloud infrastructure. For platform teams and ITDMs facing rising inference costs and GPU scarcity, this is a consolidation play worth watching.

Upbound Model Plane: Fleet-Wide AI Inference Orchestration Read More »

Nutanix NKP 2.19: Dual Native and AI Platform Strategy

Nutanix NKP 2.19: Dual Native and AI Platform Strategy

Nutanix has released NKP 2.19, introducing Dual Native infrastructure that lets organizations run virtualized and bare metal Kubernetes workloads on the same hardware. The release also adds a curated AI tooling catalog including Kubeflow, JupyterHub, and Milvus, alongside CNCF AI support certification. ECI Research examines how this positions Nutanix beyond VMware displacement toward a longer-term AI platform play.

Nutanix NKP 2.19: Dual Native and AI Platform Strategy Read More »

Vultr's Full-Stack AI Infrastructure Strategy at KubeCon 2026

Vultr’s Full-Stack AI Infrastructure Strategy at KubeCon 2026

At KubeCon 2026, Vultr outlined its full-stack AI infrastructure strategy, announcing a 20-megawatt AMD GPU supercluster in India and making the case that enterprise agentic AI workloads require more than bare-metal GPU capacity. The company’s composable stack model, regional data sovereignty architecture, and dual-vendor GPU strategy position it as a credible alternative to both hyperscalers and narrowly focused neo cloud providers. Platform engineering teams, Vultr argues, will be the organizational linchpin of enterprise AI adoption over the next two years.

Vultr’s Full-Stack AI Infrastructure Strategy at KubeCon 2026 Read More »

Dynatrace at KubeCon 2026: Agentic Deployment Meets Observability

Dynatrace at KubeCon 2026: Agentic Deployment Meets Observability

Dynatrace used KubeCon 2026 to articulate a new architecture for AI-era software delivery, one where observability data drives autonomous deployment decisions through feature flags. The company’s OpenFeature bet and recent acquisitions point to a deliberate play for the agentic deployment control plane. ECI Research data suggests developer velocity and integration ease are the top selection criteria once baseline compliance is satisfied, a tailwind for Dynatrace’s open-standard approach.

Dynatrace at KubeCon 2026: Agentic Deployment Meets Observability Read More »

Broadcom Moves VM API to CNCF: What It Means for Kubernetes

Broadcom Moves VM API to CNCF: What It Means for Kubernetes

Broadcom has donated its VM API project to the CNCF Sandbox, positioning virtual machines as first-class citizens in Kubernetes alongside containers. The move targets the growing demand for unified infrastructure control planes driven by AI and agentic workloads. ECI Research data shows vendor lock-in and integration ease are top selection criteria for platform teams evaluating modern runtimes.

Broadcom Moves VM API to CNCF: What It Means for Kubernetes Read More »

Dapr 1.19 Cuts AI Workflow Latency and Simplifies the Control Plane

Dapr 1.19 Cuts AI Workflow Latency and Simplifies the Control Plane

Diagrid has released Dapr 1.19, bringing delta-history transmission and a WorkflowsFastPath mode that cuts median workflow step latency from 553ms to 153ms. The release also consolidates two overlapping control-plane services into one, reducing operational complexity for teams running durable AI agent workloads. For government and regulated-sector operators, the combination of immutable workflow history and CNCF governance makes Dapr an increasingly credible infrastructure choice.

Dapr 1.19 Cuts AI Workflow Latency and Simplifies the Control Plane Read More »

DataAgent Fills the Kubernetes AI Detection Gap with Open-Source Pack

DataAgent Fills the Kubernetes AI Detection Gap with Open-Source Pack

DataAgent has released an open-source Detection Pack for K8sGPT that extends the CNCF Sandbox project with five new detector classes covering metrics trends, audit logs, ConfigMap changes, and pod log patterns. The pack writes to K8sGPT’s existing Result contract, requiring no integration changes from downstream tooling. It’s a technically credible move that addresses a structural blind spot in AI-assisted Kubernetes operations.

DataAgent Fills the Kubernetes AI Detection Gap with Open-Source Pack Read More »

CNCF Platinum Members Signal Open AI Infrastructure Shift

CNCF Platinum Members Signal Open AI Infrastructure Shift

Four major organizations elevated their CNCF membership to Platinum at KubeCon 2026, signaling a coordinated push to shape open AI infrastructure standards. From Adyen’s fintech-scale Kubernetes operations to CoreWeave’s AI conformance work, the moves reflect deepening anxiety about vendor lock-in and a race to define what production AI on Kubernetes actually means. ECI Research data shows that lock-in concern is nearly universal across enterprise buyers, making these governance moves more than symbolic.

CNCF Platinum Members Signal Open AI Infrastructure Shift Read More »

Platform Engineering Shapes AI Infrastructure Strategy | ECI Research

Platform Engineering Shapes AI Infrastructure Strategy | ECI Research

The Q3 2026 CNCF Technology Radar reveals that platform engineering structure, not tooling selection, is the primary driver of AI infrastructure strategy. Organizations with centralized internal developer platforms converge on deliberate approaches, while distributed teams drift into fragmented hybrid arrangements. ECI Research examines the implications for compliance, observability, and cloud native standardization.

Platform Engineering Shapes AI Infrastructure Strategy | ECI Research Read More »

CNCF Launches CKNE: Kubernetes Network Certification for AI Era

CNCF Launches CKNE: Kubernetes Network Certification for AI Era

CNCF has launched the Certified Kubernetes Network Engineer (CKNE), a performance-based credential covering CNI, traffic management, network security, and observability. The certification addresses a real skills gap as Kubernetes networking complexity grows with AI workloads. CNCF also plans two AI-focused credentials for 2027: KAIA and CKIE.

CNCF Launches CKNE: Kubernetes Network Certification for AI Era Read More »