KubeCon EU 2026 — 7 Talks We Can't Miss in Amsterdam

KubeCon + CloudNativeCon Europe 2026 runs March 23–26 in Amsterdam, and we’ll be there. Here are the sessions we’re most looking forward to — and why they matter for teams running production Kubernetes today.

🔒 Least-Privilege for AI: Authorizing Agents and MCP Tools

Luc Chmielowski (Nirmata) & Nina Polshakova (Solo.io)

As AI agents move from demos to production, the security model hasn’t caught up. This talk covers Agentgateway and Kyverno for enforcing least-privilege on MCP tool calls. If you’re running agentic AI anywhere near your infrastructure, this is mandatory viewing.

Why it matters: Every MCP tool invocation is a potential blast radius. Policy-as-code for agents isn’t optional anymore.

⚡ 1000 Services, 1 Year, 0 Downtime — Airbnb’s Zonal Cluster Migration

Sunny Beatteay (Airbnb)

Airbnb migrated 1000 services across zonal clusters over a full year with zero downtime. This is the kind of operational war story you can’t find in documentation. Expect real numbers on traffic shifting, rollback strategies, and the coordination challenges nobody warns you about.

Why it matters: Most migration guides stop at “drain and shift.” Airbnb’s scale forces decisions that expose every gap in your migration playbook.

🏗️ Helm 4 Is Here. So, Now What?

Andrew Block (Red Hat), Scott Rigby (Replicated) & Robert Sirchia (SUSE)

Helm 4 is the most significant release since Tiller was removed. New templating engine, dependency resolution changes, and the question everyone’s asking: what breaks? The maintainers themselves walk through the migration path.

Why it matters: Helm is still the de facto package manager for Kubernetes. If your deployment pipeline depends on it — and it probably does — you need to know what’s changing.

🔄 Gateway API: Bridging the Gap from Ingress to the Future

Nick Young & James Strong (Isovalent at Cisco), Katarzyna Łach & Rostislav Bobrovsky (Google), Norwin Schnyder (Airlock)

Gateway API is no longer experimental. With Ingress NGINX retiring its controller name, the migration pressure is real. This session covers the practical path from Ingress to Gateway API with real implementation experiences from multiple vendors.

Why it matters: If you’re still on Ingress, this is your migration roadmap. If you’ve already started, you’ll learn what others got wrong.

🦀 Rust vs. Go: Building a Container Network Stack From Scratch

Matt Heon (Red Hat) & Shivang K Raghuvanshi (Podman Container Tools)

A head-to-head comparison of building the same container network stack in both Rust and Go. Performance benchmarks, developer experience, memory safety trade-offs, and when each language makes sense for systems-level cloud-native work.

Why it matters: As Go developers, we respect the language’s dominance in cloud native. But knowing where Rust outperforms — and where it doesn’t — makes us better engineers.

🤖 Enterprise-Scale Migrations Using Agentic Workflows with Human-in-the-Loop

Alvaro Saurin & Jose M Navarro (Adobe)

Adobe shares how they use AI agents with human oversight to execute large-scale platform migrations. Not toy demos — actual enterprise migrations where agents propose changes and humans approve or correct. The human-in-the-loop pattern matters more than full autonomy.

Why it matters: This is where AI in platform engineering gets real. Not replacing engineers, but multiplying their capacity for tedious, high-stakes migration work.

📊 The Future of Kubernetes Scalability — Challenges of GigaWatt Computing

Maciek Różacki (Google Cloud) & Artur Rodrigues (Anthropic)

Google Cloud and Anthropic discuss scaling Kubernetes for AI workloads that consume gigawatts of computing power. Scheduling, resource management, and the fundamental scalability challenges that emerge when your cluster exists to serve foundation models.

Why it matters: AI workload scheduling is pushing Kubernetes to its architectural limits. The solutions being developed here will trickle down to every large cluster operator.

Honorable Mentions

A few more we’ll try to catch between hallway conversations:

“We Deleted Our Observability Stack and Rebuilt It With OTEL” — DigitalOcean went from 12 engineers to 4 across 20K+ clusters. That’s a compelling ROI story for OpenTelemetry.
“5000 Kubernetes Clusters, 5 Minutes” — Walmart’s edge deployment strategy with Argo. Scale numbers that make you rethink your CD pipeline.
“Dapr in the AI Era” — Diagrid on orchestrating multi-agent workflows with automatic recovery. Dapr keeps evolving in interesting directions.
“Crossplane: The Cloud Native Framework for Platform Engineering” — Upbound on why Crossplane is becoming the control plane for platform teams.

See You There

We’ll be walking the floor, attending sessions, and grabbing coffee between talks. If you’re at KubeCon EU and want to chat about cloud-native engineering, Kubernetes operations, or anything we cover on this blog — find us.

KubeCon + CloudNativeCon Europe 2026 📍 Amsterdam, The Netherlands 📅 March 23–26, 2026 🔗 Full Schedule