Cisco Live EMEA 2026: what really matters for your infrastructure
Field insights from the CNS Team in Amsterdam.
Five of our CNS consultants were on-site at Cisco Live EMEA 2026 in Amsterdam.
Our mission was to analyse the event through business, technical and strategic lenses. And to deliver an independent and expert-level perspective.
Here's what stands out for your infrastructures.
INSIGHTS FROM THE FIELD.
What we observed.
AI is an infrastructure topic, not just a model one.
Cisco's message is unambiguous.
With the right infrastructure, AI simply... doesn't scale.
The new G300 ASIC, AI-dedicated fabric architectures deployable in minutes through Nexus Dashboard, RoCE v2 to optimise GPU-to-GPU communication...
Everything converges toward the one and only goal, maximising GPU utilisation and reducing compute time.
We saw concrete results demonstrated during the week like a +20% network performance with proper configuration, translating into several days saved on a model-training cycle.
What does it change for you?
Datacenter RFPs are already integrating the notion of AI-ready DC.
Yes and if it's not yet in your requirements, it will be soon.
New Security Paradigm.
With Isovalent, Cisco bring eBPF to the core of its platforms.
Security becomes native, embedded directly into the network and OS, enabling granular and dynamic protection for distributed architectures.
Cisco is also hinting at Live Patching capabilities powered by eBPF, enabling real-time kernel patching without requiring a reboot or service interruption. If this becomes a reality, it would represent a true game changer for operations teams.
AI agents are no longer a concept, they're operational.
Every exhibitor showcased their AI agent.
We’re no longer talking about chatbots, but agents capable of taking decisions and acting on infrastructure through protocols like MCP.
Our team even built a functional agent on-site in just a few hours.
But the risks are real: prompt injection, excessive permissions, unsupervised interactions, instability. Cisco’s answer: AI‑Aware SASE and reinforced supervision of agent‑to‑infrastructure interactions.
Our field view: the agentic era has begun.
The components exist, the demos are convincing. The real question for enterprises is their infrastructure and security maturity to support these new usages.
Observability becomes mandatory.
AI in production remains a black box.
Cisco clearly understands this: Splunk is everywhere for AI infrastructure monitoring, with dashboards dedicated to inter‑agent flows, LLM execution times, and token consumption.
On the network side, Nexus Dashboard 4.2 introduces automatic anomaly detection and real‑time GPU metrics.
Observability is no longer a nice‑to‑have, it’s the cockpit.
Without it, enterprise AI remains an uncontrolled gamble.
Cisco and NVIDIA.
Cisco is no longer “just” selling switching.
The collaboration with NVIDIA results in validated reference architectures. Cisco is building a full enterprise AI ecosystem, not just hardware.
A credible strategic pivot.
And in real production?
Agentic era, embedded AI, real‑time observability:
All of this only matters if it runs in production, in your environment.
This is where CNS steps in: turning Cisco Live announcements into tangible value for your infrastructures.
The real blockers we see in the field.
Technical debt.
VLANs stacked over the past 10 years, segmentation defined “on paper” but never applied, and firewalls left in passthrough mode due to lack of visibility.
Fear of making changes in production.
Migrating a datacenter core means touching the organisation’s nervous system. Without a clear method, no one dares to press the button.
Lack of internal expertise.
On modern architectures: VXLAN, fabrics, intent‑based networking… Teams can operate the current environment, but not design the target one.
Lack of observability.
It’s impossible to validate a migration with confidence if you can’t measure the state before, during, and after.
CNS Method.
How CNS removes these blockers:
We need to act across 3 pillars.
Turning Cisco’s vision into operational reality requires three concrete pillars:
Cisco’s announcements are powerful. But between a polished demo in Amsterdam and a production datacenter with 15 years of legacy, the gap is huge.
This is exactly the gap we bridge every day.
1 ╸Designing architectures.
Designs that integrates new components (VXLAN/EVPN, Nexus Dashboard, eBPF, Splunk observability, AI agents…) into the client’s existing environment — dynamically, scalably, and without starting from scratch.
2 ╸Planning and executing migrations.
It's often in production, with no service interruption: staged migration, lab validation, progressive rollou
3 ╸Upskilling teams and automating operations.
(Network as Code, CI/CD pipelines)... while equipping monitoring so the infrastructure remains fully controlled after our intervention.
Why choosing us?
Our positioning is unique.
With a independent advisory + end‑to‑end technical expertise. We don’t sell hardware or licences, we deliver outcomes.
▶ Architecture and design.
Our consultant design datacenter, campus and hybrid cloud architectures, mastering both HLD and LLD.
Instead of handing over slide decks, we deliver deployable architectures.
▶ A recognised leader in network automation.
CNS is a recognised leader in network automation: Ansible, Terraform, GitLab pipelines...
We industrialise operations, so every change is reproducible, auditable and reversible.
▶ Observability.
We never deploy infrastructure without making it observable: Splunk, Grafana, Dynatrace, OpenTelemetry...
Are all aligned with business needs and existing tools.
▶ Environments labs.
Before every critical migration, we simulated and test scenarios in the Lab. It enables confident production cutovers.
▶ Skills and post-project autonomy.
Our internal trainings are built by field experts and adapted for clients to ensure post-project autonomy.
Do you want to discuss