This is the narrative companion to a talk-by-talk deep-dive series on KubeCon + CloudNativeCon India 2026, held in Mumbai on June 18–19. One field report, both days: what each day felt like, what the keynotes signalled about where cloud-native is heading in India, and the themes that tied the sessions together. The full deck-by-deck index — all 29 deep dives — lives on the series page; this post is the story around them.
KubeCon came to Mumbai for the first time. India's cloud-native story isn't about adopting Kubernetes anymore — that's mostly settled. The questions on stage were the next ones: how to run it at population scale, keep it secure under regulators, pay for the GPUs, and keep platform complexity manageable. Day 1 set the direction; Day 2 was about running it in production.
THE DEEP DIVESEvery talk, read slowly → the series index
17 Day-1 deep dives and 12 Day-2 deep dives, grouped by track, each unpacking the slides, architecture, and trade-offs.
Day 1 — how it actually felt
Registration opened at 8:00am; the keynotes started at 9:45am in the main hall (Jasmine 2), run as a series of short talks rather than one long session.
The focus was scale and sovereignty from the start. The opening talks were about population-scale systems — UPI payments, cricket streaming, national identity — not developer-experience demos. The recurring point: in India, cloud-native runs critical national infrastructure.
Midday was the usual mix — Project Pavilion, the Solutions Showcase, lightning talks, and hallway conversations. The afternoon split into parallel tracks: platform engineering, security and policy, AI/agents, and observability/storage/edge.
Fig 1 — the arc of day 1: keynote relay, midday community, four parallel afternoon themes.
Day 1 keynotes — what the morning signalled
After the welcome and opening remarks from Jonathan Bryce and Chris Aniszczyk of the Linux Foundation, the keynote relay made the day's thesis obvious. A few that set the tone:
- Sovereign AI (Avi Kothari & Vinayak Gavariya, Sarvam) — building AI capability that a country actually controls, on infrastructure it owns. The "sovereignty" theme recurred all day.
- Chaos as a platform (Aditya Sridasyam, Flipkart & Uma Mukkara, Harness) — chaos engineering productised into a self-serve platform, not a one-off game day.
- LLMs on Kubernetes (Shrinidhi Venkataraman & Nithin R, AstraZeneca) — running large models as first-class Kubernetes workloads, in a regulated pharma context no less.
- AI Factories (Saiyam Pathak, vCluster) — the "factory" framing for standardised, repeatable AI infrastructure.
- Population-scale AI at NPCI (Tittu Varghese) — the organisation behind UPI talking about AI at the scale of a nation's payments.
- Cloud Native at scale (Rapido) and JioHotstar cricket streaming (JioStar, Pradeep Bishnoi) — two of the most extreme real-time scaling stories you'll hear anywhere.
The through-line: AI workloads are now the gravity well the rest of the platform bends around, and scale + sovereignty + cost are the constraints everyone's optimising against. Hold that lens up to the afternoon talks and they snap into focus — almost every one is a different answer to "how do we keep this manageable as it grows."
Day 1 themes
The four threads that ran through Day 1's technical sessions — and which the deep dives are grouped by:
| Theme | The recurring question |
|---|---|
| Platform & app delivery | How do we give hundreds of teams a paved road without every team rebuilding the platform — and without one giant shared cluster becoming a blast radius? |
| Security, policy & identity | How do we get strong guarantees (policy-as-code, least privilege, safe disclosure, federated auth) without slowing delivery to a crawl? |
| AI, agents, GPU & serving | How do we orchestrate agentic systems in production, observe what they actually did, serve models that scale, and not set GPU budget on fire? |
| Observability, edge & storage | How do we see what's happening (cheaply), push compute to the edge, and run stateful storage on Kubernetes without fear? |
Day 2 — how it actually felt
Day 2 was thinner than Day 1 — some people leave early — but the ones who stayed were there for the deep technical tracks. Registration and the cloakroom opened early in the Lotus Foyer; the keynotes started 9:45am in Jasmine 2, again as a series of short talks.
If Day 1 was about scale and sovereignty, Day 2 was about running it in production. The keynotes were operational: 911 workloads on a shared cluster, RBI compliance as code, an NBFC becoming a full bank on Kubernetes, SBI's sovereign cloud. The afternoon tracks were hands-on — debugging 60 idle GPUs, ditching kube-proxy at 10M sessions/day, two-node edge storage, sandboxing LLMs, zero-trust fintech with Cilium. Less about the future, more about specific bugs and fixes.
The keynote line-up was mostly Indian companies. Motorola Solutions on emergency-call workloads, Nirmata + Razorpay on RBI compliance-as-code, Slice going from NBFC to bank, SBI on a sovereign cloud, a panel on the Indian banking sector's cloud-native shift. The pattern: regulated, high-stakes, population-scale systems, increasingly built here.
Fig 2 — the arc of day 2: a banking-heavy keynote relay, midday community, then deep parallel afternoon tracks.
Day 2 keynotes — what the morning signalled
After "Welcome Back" and opening remarks, the relay leaned hard into regulated, high-stakes infrastructure. The ones that set the tone:
- From Kubernetes Learning to Community Leadership (Lavanya Anbalagan, MatrixKube) — a pipeline-of-people message: how the next generation of cloud-native engineers gets built, and why community leadership matters as much as code.
- Running 911 & Mission-Critical Workloads on Shared Kubernetes Platforms (Manoj K R, Motorola Solutions) — when a dropped packet is a dropped emergency call, how do you safely co-tenant life-critical workloads on a shared platform?
- The Human Cost of Security in the AI Era (Sonali Srivastava, Improving) — a deliberately non-technical keynote about burnout, alert fatigue, and the people behind the dashboards as AI multiplies the security surface.
- Automating RBI Compliance With Unified Policy-as-Code (Jim Bugwadia, Nirmata & Ashwath Kumar, Razorpay) — turning Reserve Bank of India regulatory requirements into Kyverno-style policy that's enforced automatically rather than audited after the fact.
- The Vertically Integrated Bank (Faeka Ansari, Slice Financial Bank) — scaling from NBFC to a full integrated bank on Kubernetes, developer-experience first.
- SBI Meghdoot — Modern Sovereign Cloud (Rajesh Singh, SBI) — the State Bank of India's own cloud, serving a customer base the size of a large country, built for sovereignty.
- One Deploy Plane, Every Runtime (Aqeel Nazeer, Shopify) — how Shopify generalised its cloud-native deploy plane beyond Kubernetes, so one control surface drives many runtimes.
The through-line from Day 1's "AI is the gravity well" shifted to something more grounded: regulation and trust are the constraints that actually shape these systems. Policy-as-code, sovereignty, multi-tenancy under life-or-death SLAs, and the human toll of running it all — that's the closing day's thesis.
Day 2 themes
| Theme | The recurring question |
|---|---|
| AI serving, gateways & agents | How do we put LLMs and autonomous agents into production safely — gateways in front, GPUs found and shared, inference observed, agents isolated and sandboxed? |
| Networking & the kernel | How do we make Kubernetes networking fast and debuggable at real scale — past kube-proxy, into eBPF/Cilium, air-gapped, and down to netlink? |
| Platform, DevEx & supply chain | How do we make the platform invisible (Crossplane/Score), the builds lean and secure (buildpacks), and the developer path frictionless? |
| Core K8s, edge, storage & security | How do we keep the control plane healthy (etcd, API server, resiliency), run storage at the two-node edge, and hold a zero-trust line in fintech? |
Takeaways across both days
- AI is the new gravity. Sovereignty, GPUs, serving, and agents dominated Day 1 — the platform exists to feed the models now.
- Scale is assumed. The interesting questions weren't "can Kubernetes do this" but "how do we do it at population scale, under regulators, without burning the budget."
- Day 2 made it operational. Regulation became a first-class design input — RBI compliance-as-code, sovereign clouds, 911-grade multi-tenancy were keynote material, not edge cases.
- The AI conversation got concrete. Less "agents will change everything," more gateways, GPU queueing, inference metrics, sandboxes, and isolation.
- Networking and the platform went deep. Past kube-proxy, into Cilium/eBPF, air-gap, netlink; and a steady push to make the platform disappear — Crossplane/Score, buildpacks, a CNCF DevEx roadmap.
- India is setting the agenda. Real-time, huge, sovereign, cost-sensitive — banking, payments, and public-scale systems anchored both days, the centre of gravity rather than a regional sidebar.
FAQ
Is this the official schedule or your notes?
It's my field report and reading of the talks I attended, anchored to the published schedule. The keynote line-ups follow the official agenda; the deep dives (on the series page) are my own expanded write-ups of each session's slides, not official transcripts.
Where's the per-talk index?
On the series page — all 29 deep dives, grouped by track across both days, each linking to a full deck-by-deck write-up. This field report is the narrative; the series page is the index.
Why one report for both days instead of two?
Because the two days are one story: Day 1 set the vision (scale, sovereignty, AI as gravity) and Day 2 made it operational (regulation, networking, the platform disappearing). Reading them together is the point — the arc only makes sense end to end.
References
- cvam.sight — KubeCon Mumbai 2026 series · all 29 deep dives, grouped by track
- KubeCon + CloudNativeCon India 2026 — full schedule · sched.com
- KubeCon + CloudNativeCon India · event home (Linux Foundation)
- CNCF · the projects behind most of these talks