KubeCon EU 2025: Insights and Trends Shaping the Cloud Native Landscape

A comprehensive analysis of the cloud native community's premier event in London

April 1-4, 2025 ExCeL London, UK 10,000+ Attendees

Event Overview

KubeCon + CloudNativeCon Europe 2025 brought together over 10,000 cloud native enthusiasts, developers, operators, and stakeholders at ExCeL London from April 1-4, 2025. As the Cloud Native Computing Foundation's flagship conference, this event has become the central gathering point for the Kubernetes ecosystem and broader cloud native community.

This year's conference featured more than 200 sessions across multiple tracks, co-located events, and workshops, reflecting the rapid growth and maturation of the cloud native landscape. With a strong focus on AI integration, security advancements, and platform engineering practices, KubeCon EU 2025 highlighted how Kubernetes has evolved from a container orchestration platform to the foundation of modern application development and infrastructure management.

Conference Highlights:

  • 229 technical sessions across 8 primary tracks
  • 12 co-located events including Cloud Native & Kubernetes AI Day
  • Hands-on tutorials and workshops for practitioners
  • Project pavilion featuring 50+ CNCF projects
  • Numerous networking opportunities with the community's brightest minds

Key Themes and Trends

AI and ML Integration

AI and ML workloads dominated the conversation at KubeCon EU 2025, with a clear shift from theoretical discussions to production-grade implementations. Organizations are increasingly leveraging Kubernetes as the foundation for their AI infrastructure, citing its flexibility, scalability, and growing ecosystem of specialized tools.

The integration of large language models (LLMs) and generative AI into cloud native applications was particularly prominent, with many sessions focusing on efficient GPU resource management, model serving architectures, and scaling ML pipelines. Notable advancements included:

  • Enhanced scheduling capabilities for GPU/TPU workloads with Dynamic Resource Allocation
  • Streamlined approaches for AI model and dataset loading through image volumes
  • Production patterns for running LLMs on Kubernetes
  • Checkpoint/restore mechanisms for resilient AI workloads
"We're witnessing a paradigm shift from Kubernetes merely hosting AI workloads to becoming the standard platform for enterprise AI operations at scale."

Security Enhancements

Security remained a cornerstone of KubeCon EU 2025, with significant emphasis on supply chain security, identity-based trust models, and automated policy enforcement. The community has clearly matured in its approach to security, shifting from individual controls to comprehensive security strategies that span the entire application lifecycle.

Key security trends observed at the conference included:

  • Adoption of supply chain security frameworks like SLSA and GUAC
  • Integration of AI-specific security controls for protecting models and training data
  • Advancements in Validating Admission Policy and policy as code approaches
  • Confidential computing techniques for sensitive workloads
  • Automated vulnerability management and container image scanning
"Identity-based trust is becoming the foundation of Kubernetes security, replacing traditional perimeter-based approaches with fine-grained, context-aware access controls."

Platform Engineering Evolution

Platform engineering has evolved from a buzzword to an essential practice, with organizations sharing concrete implementations and metrics for success. The focus has shifted from building platforms to measuring developer experience and quantifying productivity improvements.

Several key patterns emerged in the platform engineering domain:

  • Internal developer platforms leveraging Kubernetes APIs through custom controllers and operators
  • Increased use of GitOps workflows for platform management
  • Golden paths that balance developer autonomy with operational requirements
  • Integration of AI assistants to improve platform usability
  • Metrics and KPIs for measuring platform effectiveness
"The platform engineering community is moving beyond technical implementation details to focus on the human aspects of developer experience and organizational impact."

Multi-Cluster and Edge Computing

As Kubernetes deployments scale across organizations, multi-cluster management and edge computing have become increasingly important themes. KubeCon EU 2025 featured numerous sessions on orchestrating resources across distributed environments, highlighting projects like KubeStellar, Karmada, and ClusterAPI.

Edge computing use cases were particularly prominent, demonstrating how cloud native technologies are extending beyond traditional data centers to support IoT, retail, telecommunications, and industrial applications. Key advances included:

  • Graduated KubeEdge project for edge computing
  • Lightweight Kubernetes distributions optimized for resource-constrained environments
  • Networking solutions for managing connectivity across distributed clusters
  • Automated lifecycle management for edge deployments

Sustainability and Green Computing

A notable emerging theme at KubeCon EU 2025 was sustainability and green computing. Several sessions addressed the environmental impact of cloud native infrastructure and proposed approaches for optimizing resource usage and reducing carbon footprints.

Highlights in this area included:

  • Green AI practices for reducing energy consumption of ML workloads
  • Resource optimization techniques for minimizing idle capacity
  • Metrics and observability tools for measuring energy efficiency
  • Case studies on sustainable infrastructure design
"The cloud native community is increasingly recognizing its responsibility to address environmental concerns through efficient resource utilization and sustainable practices."

Notable Sessions by Track

AI and ML Track Highlights

Production-Ready LLMs on Kubernetes: Patterns, Pitfalls, and Performance

Priya Samuel (Elsevier) & Luke Marsden (MLOps Consulting)

This session provided practical insights into deploying LLMs on Kubernetes, covering model serving architectures, resource optimization, and performance tuning. The speakers shared real-world patterns that have proven successful in production environments, along with common pitfalls to avoid.

Scaling To Thousands of GPUs With Ease: Multi-Region Large Model Training on Kubernetes

Yongxi Zhang, Meng Duan & Rongrong Wu (China Mobile)

This presentation showcased China Mobile's approach to large-scale AI training across multiple regions, demonstrating how they orchestrated thousands of GPUs for distributed training workloads while maintaining efficiency and reliability.

Yes You Can Run LLMs on Kubernetes

Abdel Sghiouar & Mofi Rahman (Google Cloud)

The speakers demystified the process of running large language models on Kubernetes, demonstrating that organizations of various sizes can leverage their existing Kubernetes infrastructure for AI workloads without specialized expertise.

Security Track Highlights

Trust No One: Secure Storage With Confidential Containers

Aurélien Bombo (Microsoft)

This session explored how confidential computing technologies can protect sensitive data in Kubernetes storage systems, ensuring that even platform administrators cannot access encrypted content without proper authorization.

Signed, Sealed, Delivered - Sign and Verify All the Things

Jeremy Rickard (Microsoft)

A comprehensive overview of artifact signing and verification in the Kubernetes ecosystem, covering tools like Notary v2, Sigstore, and OCI registries for implementing secure supply chain practices.

Zero Trust at Shopify Scale: Automating MTLS Across Thousands of Services

Dani Santos & Michelle Mali (Shopify)

Shopify engineers shared their journey implementing zero trust architecture with mutual TLS across their extensive service mesh, highlighting automation techniques that made this massive undertaking manageable.

Observability Highlights

The State of Prometheus and OpenTelemetry Interoperability

Arthur Sens (Grafana) & Juraj Michalek (Swiss RE)

This session explored the evolving relationship between Prometheus and OpenTelemetry, demonstrating how these two foundational observability projects can work together in modern monitoring stacks.

Taming 50 Billion Time Series: Operating Global-Scale Prometheus Deployments on Kubernetes

Orcun Berkem & Alan Protasio (AWS)

AWS engineers shared techniques for scaling Prometheus to handle extreme metric volumes, providing insights into sharding, long-term storage, and query optimization for massive observability deployments.

First Day Foresight: Anomaly Detection for Observability

Prashant Gupta & Kruthika Prasanna Simha (Apple)

This presentation demonstrated how Apple leverages ML-based anomaly detection to identify potential issues before they impact users, integrating predictive capabilities into their observability stack.

Storage and Data Management Highlights

Trino and Data Governance on Kubernetes

Sung Yun & Aki Sukegawa (Bloomberg)

Bloomberg engineers shared their approach to implementing data governance controls within Trino deployments on Kubernetes, balancing flexibility with compliance requirements for financial data.

The Future of Data on Kubernetes: From Database Management To AI Foundation

Melissa Logan, Nimisha Mehta, Gabriele Bartolini & Brian Kaufman (Various Organizations)

This panel discussion explored how data management on Kubernetes is evolving to support AI workloads, addressing challenges in data persistence, caching, and lifecycle management for AI applications.

Dancing With the Pods: Live Migration of a Database Fleet While Serving Millions of Queries

Jayme Bird & Manish Gill (ClickHouse)

This session detailed the techniques used to migrate a large-scale database deployment across Kubernetes clusters with minimal downtime, highlighting strategies for maintaining data consistency and performance during transitions.

Platform Engineering Highlights

From Metal To Apps: LinkedIn's Kubernetes-based Compute Platform

Ahmet Alp Balkan & Ronak Nathani (LinkedIn)

LinkedIn engineers shared their journey building a comprehensive platform that spans from bare metal infrastructure to application deployment, illustrating how they've standardized infrastructure management for thousands of developers.

Platform Engineering Loves Security: Shift Down To Your Platform, Not Left To Your Developers!

Maxime Coquerel (Royal Bank of Canada - RBC) & Mathieu Benoit (Humanitec)

This session advocated for embedding security controls within platform capabilities rather than relying on developer implementation, demonstrating how this approach improves both security posture and developer experience.

Starting and Scaling a Platform Engineering Team

Camille Fournier (Independent) & Ian Nowland (Junction Labs)

Industry veterans shared organizational strategies for building effective platform teams, addressing common challenges in staffing, prioritization, and measuring impact.

Keynote Highlights

The keynotes at KubeCon EU 2025 reflected the maturing cloud native ecosystem while highlighting emerging challenges and opportunities. Key announcements and themes included:

CNCF Project Updates

Industry Directions

"We're seeing Kubernetes transition from a technology platform to a business-critical infrastructure that demands mature operational practices, robust security, and sustainable design principles."

Industry Impact and Future Directions

KubeCon EU 2025 demonstrated that Kubernetes has firmly established itself as the foundation for modern application infrastructure, with organizations now focusing on building higher-level capabilities and addressing specific business challenges. Several key trends will shape the cloud native landscape in the coming year:

AI-Native Infrastructure

Kubernetes is rapidly evolving to become the default platform for AI workloads, with specialized tools and patterns emerging for different stages of the AI lifecycle. This trend will accelerate as organizations move from experimentation to production AI deployments.

Unified Security Frameworks

The fragmented landscape of cloud native security tools is consolidating around comprehensive frameworks that span the entire software lifecycle, with increasing focus on supply chain security and automated policy enforcement.

Platform as a Product

The concept of treating internal platforms as products with clear user experiences, measured outcomes, and continuous improvement cycles is gaining traction, with organizations investing in platform teams that bridge technical and business concerns.

Edge and Multi-Cluster Management

As Kubernetes deployments expand beyond centralized clusters, tools and practices for managing distributed environments are maturing, enabling new use cases in edge computing, hybrid cloud, and multi-region deployments.

WebAssembly Integration

WebAssembly (Wasm) is emerging as a complementary technology to containers, offering lightweight, secure execution environments for specific workloads and extending the capabilities of existing cloud native platforms.

Practical Takeaways

Based on the discussions and presentations at KubeCon EU 2025, here are key recommendations for organizations looking to advance their cloud native journey:

Focus Area Recommendations
AI Workloads
  • Evaluate dynamic resource allocation (DRA) for GPU management
  • Consider specialized operators like KubeFlow and Slinky for ML orchestration
  • Implement checkpoint/restore mechanisms for long-running AI tasks
Security
  • Adopt supply chain security tools like Sigstore, GUAC, and Kubescape
  • Implement validating admission policies for automated enforcement
  • Use SPIFFE/SPIRE for workload identity across environments
Observability
  • Integrate OpenTelemetry for unified observability data collection
  • Explore eBPF-based tools for deep system visibility
  • Implement SLOs and error budgets for service reliability
Platform Engineering
  • Focus on developer experience metrics to measure platform effectiveness
  • Implement internal developer portals for service discovery and self-service
  • Use CRDs and operators to extend Kubernetes for specific domains
Multi-Cluster
  • Evaluate projects like Karmada, ClusterAPI, and KubeStellar for fleet management
  • Implement centralized observability and governance across clusters
  • Standardize on GitOps workflows for consistent deployment

Conclusion

KubeCon + CloudNativeCon Europe 2025 demonstrated the continued evolution of the cloud native ecosystem, with Kubernetes firmly established as critical infrastructure for modern applications. The conference highlighted how organizations are moving beyond basic adoption to tackle complex challenges in AI, security, platform engineering, and multi-cluster management.

The maturity of the ecosystem was evident in the depth of technical discussions, with less focus on introductory content and greater emphasis on production-grade implementations, operational excellence, and sustainable practices. As Kubernetes celebrates its tenth anniversary, the community is increasingly turning its attention to the human aspects of cloud native adoption, including developer experience, organizational transformation, and inclusive governance.

Looking ahead, the integration of AI capabilities, advancement of security practices, and expansion to edge environments will continue to drive innovation in the cloud native landscape. Organizations that embrace these trends while focusing on operational excellence and developer productivity will be well-positioned to leverage Kubernetes as a foundation for digital transformation.

"KubeCon EU 2025 showcased how Kubernetes has evolved from a revolutionary technology to an essential platform that enables innovation across industries and use cases."

© 2025 Cloud Native Blog. All insights based on KubeCon + CloudNativeCon Europe 2025.