How JioHotstar Streams IPL 2026 to 65 Million Concurrent Viewers on AWS

Abhishek Gautam··8 min read

Quick summary

JioHotstar hit 65.2 million concurrent viewers during India vs England T20 2026 — a world record. Here is the AWS EKS architecture, multi-CDN strategy, and 800+ microservices powering IPL 2026.

IPL 2026 starts March 28. Over the next 65 days, JioHotstar will attempt to do something no streaming platform has reliably done before: deliver live cricket to tens of millions of simultaneous viewers in India without buffering, at scale, across 2G connections in rural Bihar and 5G connections in Mumbai, simultaneously. The platform already holds the world record — 65.2 million concurrent viewers during India vs England in the T20 World Cup 2026 semi-final. Here is exactly how they do it.

What JioHotstar Is in 2026

JioHotstar is the product of a merger that completed in February 2025: JioCinema (Reliance/Viacom18) and Disney+ Hotstar combined into a single platform under the JioStar joint venture. Reliance holds 63.16% of the venture, Disney holds 36.84%.

The result is India's largest streaming service by almost every metric: 500 million registered users, 300,000 hours of content, 19 language options, 100+ live TV channels, and the exclusive digital streaming rights to IPL through the current broadcast cycle. Star Sports Network handles television broadcast; JioHotstar handles everything digital.

For IPL 2026 specifically — which runs from March 28 through May 31 — JioHotstar has signed six digital sponsors and five TV sponsors, with subscription plans starting at ₹149 for three months. Jio prepaid users get access bundled with select plans. The platform streams IPL in 4K.

The World Record: 65.2 Million Concurrent Viewers

During the India vs England semi-final at the ICC Men's T20 World Cup 2026, JioHotstar recorded 65.2 million peak concurrent viewers — the highest concurrency ever achieved on any digital platform for any live event globally. The match generated 619 million total views, making it the most-streamed T20 International in history.

To understand what 65.2 million concurrent viewers means architecturally: imagine 65 million people each requesting a continuous video stream, each generating heartbeat signals, chat messages, reaction events, and authentication checks — all simultaneously, all expecting sub-3-second start times and uninterrupted playback.

For context: Netflix's global peak concurrent viewership for its most popular content sits around 40-50 million. The Super Bowl peaks around 10-12 million concurrent streams in the US. JioHotstar's IPL numbers are in a category of their own.

The AWS EKS Core: 800+ Microservices

At the heart of JioHotstar's architecture is a massive Amazon Elastic Kubernetes Service (EKS) cluster orchestrating over 800 microservices. Each microservice handles a specific function: authentication, live reactions, chat, dashboard rendering, recommendation engine, payment processing, subscription validation, and hundreds more.

The infrastructure specification during peak IPL load:

  • 16 TB of RAM across the cluster
  • 8,000 CPU cores
  • 32 Gbps peak data transfer speed

The EKS cluster uses a custom autoscaler built by JioHotstar's engineering team that can spin up new pods within 30 seconds of detecting rising concurrency. This is critical for cricket: viewership does not ramp up gradually during an IPL match — it spikes instantly when India's top batters come to the crease or when a wicket falls in a close finish. The autoscaler has to respond in near-real-time, not in minutes.

Standard Kubernetes Horizontal Pod Autoscaler reacts to CPU and memory metrics, which lag actual user load. JioHotstar's custom autoscaler reacts to concurrency metrics directly — the number of active streams — which gives it a 60-90 second advantage over standard HPA in responding to sudden spikes.

The Multi-CDN Strategy: 90% Cache Hit Rate

Video delivery at 65 million concurrent streams cannot come from origin servers — the bandwidth would be physically impossible. JioHotstar's solution is a multi-CDN architecture with an in-house CDN load optimizer.

The load optimizer sits in front of multiple CDN providers and dynamically routes each viewer's stream through the least congested path in real time. If Akamai is congested in Maharashtra, traffic shifts to Cloudflare or AWS CloudFront automatically, without the viewer experiencing any interruption. The system provides redundancy: if any single CDN has an outage, traffic redistributes within seconds.

During IPL streaming, over 90% of content requests are served directly from CDN cache rather than origin servers. This means JioHotstar's origin infrastructure is handling roughly 6.5 million unique stream requests at peak — still enormous, but a tenth of the apparent load.

The video encoding pipeline produces multiple renditions of every stream simultaneously: 4K, 1080p, 720p, 480p, 360p, and low-bandwidth versions for 2G connections. The client-side player uses adaptive bitrate (ABR) streaming — it continuously measures the viewer's available bandwidth and switches renditions seamlessly without interrupting playback. A viewer on a crowded stadium wifi gets 360p; a viewer on home fiber gets 4K. Same stream, same infrastructure, different rendition.

Handling India's Network Reality

India's network infrastructure is unlike any other major streaming market. The same IPL match is watched by:

  • A viewer on Jio 5G in Bangalore with 500 Mbps
  • A viewer on Airtel 4G in Pune with 30 Mbps
  • A viewer on BSNL 2G in rural Uttar Pradesh with 512 Kbps
  • A viewer on shared broadband at a tea stall in Chennai

All four viewers expect the stream to work. This is why the adaptive bitrate pipeline and CDN edge caching are not optional features — they are the core engineering problem JioHotstar exists to solve.

Jio's telecom infrastructure (450+ million subscribers) gives JioHotstar a structural advantage: Reliance's network data informs CDN placement decisions. JioHotstar can pre-position CDN edge nodes in locations where Jio's network data shows high cricket viewership density, reducing last-mile latency before the match even starts.

The Star Sports TV Side

While JioHotstar handles digital, Star Sports Network handles television broadcast across dedicated channels: Star Sports 1, Star Sports 2, Star Sports 1 Hindi, and regional language feeds in Tamil, Telugu, Kannada, and Bengali.

The TV broadcast matters for the streaming infrastructure in one specific way: TV viewers represent the viewership that is not on JioHotstar. India still has hundreds of millions of cable and DTH television subscribers. The cricket fans who watch on Star Sports are not generating concurrent stream load on JioHotstar. As India's TV-to-digital transition continues, JioHotstar's concurrent viewer numbers will only increase — the 65.2 million record is not the ceiling.

What Developers Can Learn From This Architecture

Three patterns from JioHotstar's IPL architecture are directly applicable to any high-concurrency application:

Custom autoscaling beats generic autoscaling for bursty traffic. If your application has traffic patterns that spike on external events (sports, news, product launches), build or configure an autoscaler that reacts to your actual load metric — concurrent users or active sessions — not CPU and memory, which lag by 2-5 minutes.

Multi-CDN with a smart router beats single-CDN contracts. The marginal cost of a CDN load optimizer is worth it at scale. Single-CDN contracts optimise for unit cost; multi-CDN with dynamic routing optimises for availability. For live events where one failed CDN node means millions of dropped streams, availability wins.

Cache hit rate is your primary scaling lever for video. At 90% cache hit rate, JioHotstar serves 10x the apparent load from origin infrastructure. Every percentage point improvement in cache hit rate reduces origin load proportionally. This applies to any static or semi-static content: precompute, cache aggressively, and treat cache miss as the exceptional path.

Key Takeaways

  • JioHotstar is the official digital streaming platform for IPL 2026 (March 28 – May 31), formed from the February 2025 merger of JioCinema and Disney+ Hotstar; Reliance holds 63.16%, Disney 36.84%
  • 65.2 million peak concurrent viewers during India vs England T20 World Cup 2026 semi-final — the highest concurrency ever recorded on any streaming platform globally
  • 800+ microservices on AWS EKS: 16 TB RAM, 8,000 CPU cores, 32 Gbps peak — with a custom autoscaler that spins up pods in 30 seconds
  • Multi-CDN strategy with in-house load optimizer dynamically routes viewers through least-congested CDN in real time; 90% of IPL content served from CDN cache, not origin
  • Adaptive bitrate streaming serves all of India simultaneously — from 4K on 5G to low-bandwidth on 2G — with seamless quality switching based on real-time bandwidth measurement
  • Star Sports handles TV broadcast across 6 language feeds; JioHotstar handles all digital; as India's TV-to-digital transition continues, JioHotstar's concurrent viewer records will keep climbing
  • IPL 2026 subscription from ₹149 for 3 months; Jio prepaid users get access bundled with select plans; 4K streaming available

Free Weekly Briefing

The AI & Dev Briefing

One honest email a week — what actually matters in AI and software engineering. No noise, no sponsored content. Read by developers across 30+ countries.

No spam. Unsubscribe anytime.

ShareX / TwitterLinkedIn

Written by

Abhishek Gautam

Full Stack Developer & Software Engineer based in Delhi, India. Building web applications and SaaS products with React, Next.js, Node.js, and TypeScript. 8+ projects deployed across 7+ countries.