IPL's $6.2 Billion Broadcast Deal and the AWS Infrastructure Behind It

Abhishek Gautam··8 min read

Quick summary

IPL 2023-27 rights sold for ₹48,390 crore ($6.2B) — the world's 2nd most valuable sports property per match after the NFL. Here is how JioStar built the AWS infrastructure to justify every rupee.

The IPL is the second most valuable sports property in the world per match — behind only the NFL. The ₹48,390 crore ($6.2 billion) rights deal for 2023-2027 valued each IPL match at ₹118 crore ($14.61 million). The Super Bowl earns $37 million per game. No other cricket tournament, Premier League match, or Champions League fixture comes close to this per-match valuation. And behind every rupee of that deal is an AWS infrastructure that has to perform flawlessly in front of 65 million simultaneous viewers.

The $6.2 Billion Deal: How It Broke Down

In June 2022, the BCCI auctioned IPL media rights for the 2023-2027 cycle across four separate packages. The final numbers:

Package A — TV rights (India): Won by Disney Star for ₹23,575 crore. This covers linear television broadcast on Star Sports across all language feeds — English, Hindi, Tamil, Telugu, Kannada, Bengali.

Package B — Digital rights (India): Won by Viacom18 for ₹23,758 crore. This was the historic moment: for the first time ever, digital rights for an Indian sports property sold for more than TV rights. Viacom18 at the time was Reliance's media arm, operating JioCinema.

Package C — Selected matches for non-exclusive digital: ₹294 crore.

Package D — International rights: ₹1,058 crore split across multiple regional broadcasters.

Total: ₹48,390 crore ($6.2 billion at the time of the deal). For perspective, the previous IPL rights cycle (2018-2022) sold for ₹16,347 crore. The 2023-2027 deal was a 196% increase in five years.

The per-match value of ₹118 crore means the BCCI earns more per IPL match than the Premier League earns per game ($11.2 million), the NBA earns per game ($2.6 million), or the Indian cricket board earns from bilateral Test series.

The Corporate Consolidation That Changed Everything

The $6.2 billion deal was struck between two separate companies — Disney Star and Viacom18 — who then became one company.

In November 2024, Disney completed a merger of its entire Indian media and streaming operations with Reliance's Viacom18. The deal was valued at ₹70,352 crore. The joint venture was named JioStar: Reliance holds 63.16%, Disney holds 36.84%.

In February 2025, the two streaming apps — Disney+ Hotstar and JioCinema — merged into a single platform: JioHotstar. The result: JioStar now holds both the TV rights (via Star Sports) and the digital rights (via JioHotstar) to IPL under a single ownership structure.

This means the $6.2 billion IPL deal, originally split competitively between two rival companies, is now controlled by one entity. Every IPL rupee — TV ad revenue, digital subscription revenue, streaming ad revenue — flows into JioStar.

AWS: The Technical Foundation

JioHotstar does not run its streaming infrastructure on private data centers. It runs on Amazon Web Services, specifically Amazon EKS (Elastic Kubernetes Service), with a cluster architecture that scales to the specific demands of live cricket at hundreds-of-millions-of-viewer scale.

Here is what the AWS infrastructure looks like during a live IPL match:

Compute layer: 500+ AWS CPU instances, predominantly C4.4XLarge and C4.8XLarge instance types — compute-optimised instances designed for high-throughput, CPU-intensive workloads. During peak IPL matches, the cluster scales to over 4,000 Kubernetes worker nodes.

Kubernetes orchestration: 800+ microservices run across the EKS cluster, each handling a discrete function. Authentication verifies that a subscriber's session token is valid before streaming begins. The recommendation engine selects which pre-match content appears on each user's home screen. The chat service handles real-time fan reactions. The dashboard service renders live scorecards. Each microservice scales independently based on its own load metrics.

Custom autoscaler: JioHotstar built a proprietary autoscaler on top of EKS that triggers pod scaling based on active concurrent streams — not the standard CPU/memory metrics that Kubernetes HPA uses. This matters because CPU load on a streaming pod is not proportional to concurrent viewers in the same way it is for compute-intensive workloads. The custom autoscaler reacts to the actual load metric (streams) and spins up new pods in under 30 seconds of detecting rising concurrency.

Storage and data pipeline: AWS S3 stores all historical match footage, Hawk-Eye ball-tracking data, and pre-encoded video segments. Amazon Kinesis handles real-time data streaming — the pipeline that moves live video from broadcast trucks to the encoding farm to CDN edge nodes with end-to-end latency under 10 seconds.

The Video Encoding Pipeline: From Broadcast Truck to Your Phone

Most people have no mental model of what happens between a ball being bowled at the Wankhede Stadium and that ball appearing on a phone screen in Hyderabad 8 seconds later. Here is the pipeline:

Step 1 — Capture: Multiple high-definition cameras at the ground feed raw video to an OB (Outside Broadcast) truck parked outside the stadium. The OB truck aggregates all camera feeds and produces a single broadcast stream — the same one that goes to Star Sports TV and to JioHotstar.

Step 2 — Ingest: The broadcast stream is transmitted from the OB truck via satellite uplink and fibre connection to JioHotstar's ingest servers running on AWS. The ingest layer receives the raw broadcast-quality stream (typically 50-100 Mbps uncompressed).

Step 3 — Encoding farm: AWS Elemental MediaLive transcodes the raw stream into multiple renditions simultaneously: 4K (15 Mbps), 1080p (8 Mbps), 720p (4 Mbps), 480p (2 Mbps), 360p (1 Mbps), and 240p (500 Kbps) for very low bandwidth connections. Each rendition is packaged into HLS (HTTP Live Streaming) segments — short 2-4 second chunks that can be independently requested and cached.

Step 4 — CDN distribution: The HLS segments are pushed from AWS origin servers to edge nodes across JioHotstar's multi-CDN network (Akamai, Cloudflare, AWS CloudFront, and Jio's own MEC nodes). During IPL, 90%+ of segment requests are served from CDN cache — the origin servers handle only cache misses and the first segment of a new stream.

Step 5 — Adaptive playback: The JioHotstar app on the viewer's device uses ABR (Adaptive Bitrate) logic to request the appropriate quality rendition based on the current measured bandwidth. The player continuously monitors download speed and switches renditions mid-stream if conditions change — upward if bandwidth improves, downward if it drops. The switch is seamless: the next HLS segment loads at the new quality level without buffering.

End-to-end latency target: Under 10 seconds from camera to screen. Glass-to-glass (camera lens to display) latency of under 10 seconds is the industry benchmark for live sports streaming. JioHotstar achieved 99.995% uptime with sub-10 second latency for IPL 2025 at 41 million concurrent viewers.

Jio's Network as an Unfair Advantage

JioHotstar is not just an AWS customer. It is part of Reliance, which operates India's largest telecom network with 450+ million subscribers. This gives JioHotstar a structural infrastructure advantage that no other streaming platform in India can replicate.

Jio operates Mobile Edge Computing (MEC) servers co-located with its 5G base stations across India. These MEC servers function as ultra-low-latency CDN nodes — positioned physically inside the telecom network rather than at internet exchange points. For a Jio subscriber watching IPL on a Jio 5G connection, video segments can be served from a MEC node that is one network hop away from the device rather than routing through the public internet.

The practical result: Jio subscribers watching IPL on JioHotstar experience lower latency and higher quality streams than subscribers on Airtel or Vi, because JioHotstar's CDN has physical presence inside Jio's network that competitors cannot access.

The Revenue Model Behind the $6.2 Billion Bet

JioStar needs to recover ₹48,390 crore over five years — approximately ₹9,678 crore per year — just to break even on the rights cost, before any infrastructure, operations, or content costs.

The revenue streams:

Digital subscription (JioHotstar): Subscriptions starting at ₹149 for 3 months. If JioHotstar converts 30 million paying subscribers at an average ₹500/year, that is ₹1,500 crore annually from subscriptions.

Digital advertising: JioHotstar signed six digital sponsors for IPL 2026. Programmatic ad inventory during live IPL streaming — served to identified, logged-in users with Jio's consumer data — commands premium CPMs. Digital ad revenue from IPL likely exceeds ₹4,500 crore per IPL season.

TV advertising (Star Sports): Five TV sponsors for IPL 2026. Linear TV advertising during IPL is India's most expensive inventory. Star Sports TV ad revenue from IPL is estimated at several thousand crore per season.

The math is tight. JioStar paid a 196% premium over the previous rights cycle betting that India's streaming market would grow fast enough to justify it. IPL 2026's $6.2 billion price tag is also a bet that the AWS infrastructure can keep 65 million concurrent viewers watching — because every buffering event, every crash, every dropped stream is lost ad revenue and lost subscriber trust.

Key Takeaways

  • IPL 2023-2027 rights sold for ₹48,390 crore ($6.2B) — ₹118 crore per match, making it the world's 2nd most valuable sports property per match behind the NFL ($37M per game)
  • Digital rights (₹23,758 crore) exceeded TV rights (₹23,575 crore) for the first time in Indian sports history — Viacom18's bid signalled the inflection point in India's TV-to-streaming transition
  • JioStar now holds both TV and digital IPL rights following the November 2024 Reliance-Disney merger (₹70,352 crore deal) and February 2025 JioHotstar platform merger
  • AWS EKS scales to 4,000+ Kubernetes worker nodes during peak IPL matches, running 800+ microservices on 500+ C4 compute instances with a custom autoscaler that reacts in under 30 seconds
  • End-to-end pipeline: OB truck → AWS Elemental MediaLive encoding (7 simultaneous renditions) → AWS S3/Kinesis → multi-CDN distribution → adaptive bitrate playback — all in under 10 seconds
  • Jio MEC nodes inside 5G network give JioHotstar subscribers a CDN advantage that no competitor can replicate — physical presence inside the telecom network versus public internet routing
  • 99.995% uptime at 41M concurrent viewers (IPL 2025) — the target for IPL 2026 is higher; at 65.2M concurrent viewers (T20 World Cup record), the infrastructure has already proven it can handle the load

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.