OTT app development in India 2026: what a streaming build actually costs, from DRM to CDN egress

Summary. India's OTT audience reached 601.2 million people, 41.1% of the population, in the fifth edition of the Ormax OTT Audience Report published on 17 September 2025, with 148.2 million active paid subscriptions. The number that should change your architecture is a different one: connected TV users jumped 87% in a single year to 129.2 million, roughly 35-40 million homes. Big screens mean higher bitrates, and higher bitrates hit two bills at once. On AWS Elemental MediaConvert, DRM encryption, captions and multiple audio tracks are all Professional-tier features, which starts at $0.0120 per normalized minute rather than the Basic tier's $0.0075. On the delivery side, CloudFront's $1,000-a-month Premium flat-rate plan includes 50 TB of data transfer, which at a 3 Mbps HD stream is about 38,000 viewing hours a month. For an Indian OTT service that is a rounding error. Here is the arithmetic before you commit a budget.

Most OTT cost estimates you will read stop at "it depends on features". The features are not the expensive part. Encoding your catalogue once and shipping it to a million screens is.

The India numbers, and which ones to trust

Published figures for the Indian streaming market disagree, because they measure different things. Table them rather than picking one.

Two things follow for anyone scoping a build. Audience growth slowed to 10% year on year, down from the 13-14% seen in 2023 and 2024, so the land-grab phase is over and unit economics matter. And with 601.2 million viewers against 148.2 million paid subscriptions, the ratio is roughly four viewers per paid subscription, which is why almost every serious Indian OTT plan in 2026 is hybrid: ad-supported tiers carrying reach, subscriptions carrying margin.

Shailesh Kapoor, founder and CEO of Ormax Media, singled out the connected TV shift in the report's release, noting that "the sharp rise in the Connected TV audience base is particularly striking for a market long regarded as mobile-first."

That is an engineering statement disguised as a market statement. A mobile-first service can ship a 240p-to-720p ladder and get away with it. A living-room service cannot.

Where the money actually goes

An OTT build has five cost centres. Only two of them scale with success.

Teams over-invest in the first two and under-model the last two. The build is a known quantity. The bill after launch is not.

Encoding: the DRM tier trap

AWS Elemental MediaConvert charges per minute of output, multiplied by a normalization factor that depends on codec, resolution and frame rate. The pricing page, last updated 13 May 2026, sets out two tiers.

The Basic tier covers AVC, VP8 and VP9 outputs for simple web distribution. The Professional tier covers HEVC up to 8K, AV1, multi-pass encoding, accelerated transcoding and, critically, DRM encryption, captions, multiple audio tracks and ad signalling.

Read that list again. DRM encryption, captions and multiple audio tracks are all listed as unavailable on the Basic tier. For an Indian OTT service, multi-language audio is not a premium feature, it is the product. Captions are a legal and accessibility expectation. Any licensed content requires DRM. So an Indian streaming service is on the Professional tier from day one, and the Basic tier rates you see quoted in generic cost guides do not apply to you.

Published US East (Ohio) rates:

The normalization multipliers matter as much as the rate. On the Professional tier, AVC at 30 fps is 1x for SD, 2x for HD and 4x for 4K. HEVC doubles that: 2x, 4x and 8x. AV1 at HD is 7x. Audio-only output is 0.39x. Dolby Vision is a 10x add-on at HD and 20x at 4K.

Worked example. Take a 5,000-hour catalogue, which is 300,000 source minutes, encoded once with a modest ladder: two SD AVC renditions at 30 fps (1x each), three HD AVC renditions at 30 fps (2x each), and four audio-only language tracks (0.39x each). The combined multiplier is 2 + 6 + 1.56 = 9.56x, giving 2,868,000 normalized minutes.

At Professional tier rates: the first 50,000 minutes cost $600, the next 950,000 cost $9,120, and the remaining 1,868,000 cost $13,449.60. Total: about $23,170 to encode the catalogue once, roughly ₹20 lakh at an assumed ₹87 to the dollar.

Now switch that ladder to HEVC for the CTV audience and every video multiplier doubles. Add Dolby Vision at 10x per HD rendition and the encoding bill stops being a rounding error. Those published rates are for US East (Ohio) and EU (Stockholm); Asia Pacific (Mumbai) rates differ, so price your own region before you plan.

One detail that catches teams doing short-form: the minimum billing duration is 10 seconds, so a 4-second clip bills as 10. If your product is reels, model that.

CDN egress: the number that decides your business model

This is where most OTT plans quietly break.

CloudFront's flat-rate plans, per the pricing page updated 15 July 2026, run from a $0 Free plan through Pro at $15 a month and Business at $200 a month to Premium at $1,000 a month. The Premium plan includes 500 million requests, configurable up to 6 billion, and 50 TB of data transfer, configurable up to 600 TB, for a new flat-rate price.

Do the conversion. An HD stream at 3 Mbps consumes about 1.35 GB per viewing hour. So:

A flat-rate CDN plan is built for websites and applications, not for a streaming catalogue. If your OTT service works, you will be on pay-as-you-go with a negotiated committed-use agreement inside a year, and egress will be the largest single line on your infrastructure bill. Plan the commercial conversation with your CDN before launch, not after the first spike.

Three levers actually move this number, and they are all encoding decisions:

1. Per-title encoding. A talking-head interview does not need the bitrate a cricket match needs. Quality-defined variable bitrate encoding, available on both MediaConvert tiers for AVC, typically cuts average bitrate without a visible quality loss.

1. Codec choice. HEVC and AV1 cost more to encode and less to deliver. The crossover point is a function of how many times each title is watched. For a long-tail catalogue AVC usually wins; for your top 200 titles, HEVC usually pays for itself.

1. Ladder discipline. Every rendition you add is encoded for the whole catalogue and cached at every edge. Most services ship two or three renditions more than their measured device mix justifies.

The real cost is the egress, not the build. Every architecture decision should be argued against that sentence.

What a serious Indian OTT stack looks like in 2026

• Ingest and mezzanine: object storage with lifecycle rules; keep mezzanines in infrequent-access tiers after the first 90 days.

• Encoding: MediaConvert Professional tier or an equivalent, with per-title QVBR and a ladder derived from your own device analytics.

• Packaging: CMAF with HLS and DASH manifests from one set of segments, which halves storage and cache footprint against packaging each format separately.

• DRM: multi-DRM covering Widevine, FairPlay and PlayReady, because Android, Apple and Windows or smart-TV platforms each require a different system. This is the part teams most often discover late.

• Player: one player abstraction across Android, iOS, web and CTV, with quality-of-experience telemetry from day one. Startup time, rebuffer ratio and average bitrate are the three metrics that predict churn.

• Entitlements and payments: subscription state, telecom-bundle reconciliation and device limits. In India telecom and aggregator bundles are a large share of paid subscriptions, so bundle reconciliation is core, not an integration afterthought.

• Ad stack: server-side ad insertion if you run an ad tier, which requires the ad signalling that sits in the Professional encoding tier.

For teams weighing cross-platform against native for the app layer, our Flutter app development notes and the India versus US app development cost comparison set out the trade-offs we apply, including where a native player integration is worth the extra surface area.

India-specific considerations

Language is the architecture. A service targeting Hindi plus four regional languages ships four to six audio tracks and as many subtitle tracks per title. On the Professional tier each audio-only track adds 0.39x to the multiplier and each subtitle track adds manifest complexity. Budget for it at the ladder-design stage rather than discovering it during localisation.

Connected TV changed the bitrate floor. With the CTV base at 129.2 million and up 87% year on year, a top rendition of 720p is no longer defensible. That is a direct increase in both encoding multiplier and per-hour egress.

Data protection. Viewing history, watchlists and device identifiers are personal data. We design applications aligned with Digital Personal Data Protection Act 2023 requirements, including consent capture for recommendation profiling and deletion workflows that actually reach your analytics warehouse and not only your primary database. Our DPDP-ready app development notes cover the engineering pattern.

Network reality. A meaningful share of Indian viewing happens on constrained mobile networks. Low-latency startup, aggressive first-segment sizing and a working offline-download path are not nice-to-haves in this market.

Cloud spend discipline. Egress-dominated bills respond to commitment structures and architecture, not to instance right-sizing. That is a different FinOps practice from a typical SaaS workload, and we walk through the India-specific version in cutting cloud spend for Indian teams.

How we scope an OTT engagement

eCorpIT was founded in 2021 and works from Gurugram as a CMMI Level 5 assessed, MSME-certified organisation with senior-led, multi-disciplinary engineering teams, and we are partners with AWS, Microsoft and Google. On a streaming build we run four phases.

Phase 1, architecture and cost model (2-3 weeks). We take your catalogue size, target device mix and expected viewing hours, and produce the encoding ladder, packaging strategy and a modelled monthly infrastructure bill in your own region and currency. You get the numbers before the commitment, not after.

Phase 2, platform build. Ingest, encoding pipeline, packaging, multi-DRM, entitlements and CMS. Automated regression on the pipeline itself, because a broken re-encode is discovered by viewers otherwise.

Phase 3, apps. Android, iOS, web and connected TV with a shared player abstraction and QoE telemetry wired in from the first build, not retrofitted.

Phase 4, operate and optimise. Ladder tuning against measured device mix, codec migration for high-traffic titles, and CDN commercial optimisation. This is where the modelled savings turn real, and it is ongoing work rather than a handover.

Who this is for: broadcasters and media businesses moving direct-to-consumer, education and fitness businesses with large video catalogues, and enterprises running internal video platforms at scale. Who it is not for: a business with under 100 hours of content, which should use an off-the-shelf video platform and revisit this in two years.

FAQ

How eCorpIT can help

eCorpIT builds streaming platforms and the apps in front of them for media, education and enterprise clients from Gurugram, and we start every engagement with the cost model rather than the wireframes. Our senior engineering teams size the encoding ladder against your real device mix, set up multi-DRM packaging correctly the first time, and build Android, iOS, web and connected TV clients on a shared player with quality-of-experience telemetry from day one. We design applications aligned with Digital Personal Data Protection Act 2023 requirements for viewing history and profiling data. If you are planning an OTT launch or your current egress bill has stopped making sense, talk to us and we will model your numbers before anyone writes code.

References

1. India has 601 million OTT users and 148 million active paid subscriptions, reveals Ormax's research report — India Brand Equity Foundation, 17 September 2025.

1. AWS Elemental MediaConvert Pricing — Amazon Web Services, page updated 13 May 2026.

1. Amazon CloudFront CDN pricing — Amazon Web Services, page updated 15 July 2026.

1. CloudFront pay-as-you-go pricing — Amazon Web Services.

1. CloudFront flat-rate pricing plans — AWS Documentation.

1. AWS Elemental MediaPackage pricing — Amazon Web Services.

1. Amazon S3 pricing — Amazon Web Services.

1. Working with reserved queues in MediaConvert — AWS Documentation.

1. AWS Elemental MediaConvert features — Amazon Web Services.

1. India's Streaming Universe Grows to 547 Million Users as Paid Subscriptions Stagnate — Variety.

1. India has 601 million OTT users and 148 million active paid subscriptions: Ormax's research report — MediaNews4U.

1. India's OTT audience hits 601 million, paid subscriptions at 148 million: Ormax report — afaqs!

1. OTT Market in India Size, Trends, Share and Report — IMARC Group.

1. Streaming insights — Ormax Media.

Last updated: 22 July 2026.