Cloud Egress Fees in India 2026: Why Data Transfer Costs Hurt

Last month I sat with the finance head of a mid-sized D2C brand in Bengaluru who couldn't figure out why their AWS bill had crept from ₹2.8 lakh to ₹4.1 lakh in six months. Compute was flat. Storage had barely moved. When we pulled the Cost Explorer report and filtered by usage type, the culprit was staring at us: DataTransfer-Out-Bytes and inter-region transfer charges accounted for nearly ₹95,000 that month. Nobody on the team had ever looked at that line.

This is the quiet tax of cloud computing. You pay to put your data in, you pay to store it, and then you pay again every time it leaves. Egress fees are the most misunderstood, least budgeted, and most avoidable part of cloud spending. And they hit Indian businesses harder than most people realise, because our architectures often span the Mumbai and Singapore regions, feed CDNs serving customers across the country, and shuttle backups between accounts.

In this post I'll break down exactly how cloud data transfer costs India businesses face are structured across AWS, GCP and Azure, walk through a real optimisation I did for a logistics client, give you a comparison table, a step-by-step audit process, and a FAQ that answers the questions I get asked most. By the end you'll know where the money leaks and how to plug it.

Key Takeaways

• Egress (data transfer out to the internet) is typically the costliest transfer category, often ₹7–₹9 per GB on major clouds after the small free tier.
• Inter-region and inter-AZ transfers are silent budget killers — moving data between Mumbai and Singapore can cost as much as egress to the internet.
• Data into the cloud is almost always free; the pain is always on the way out and sideways.
• A CDN in front of your app can cut egress by 60–80% for content-heavy workloads and is usually cheaper per GB than raw egress.
• Architecting for a single region, using private endpoints, and compressing payloads are the highest-ROI fixes most SMBs never implement.
• Always model egress before migration — it's the number that turns a "cheaper on cloud" business case negative.

What exactly are cloud egress and data transfer fees?

Cloud providers charge for data movement in three broad ways, and understanding the difference is half the battle.

Egress to the internet is what most people mean by "egress." Any byte that leaves the cloud provider's network and travels to a user's browser, a mobile app, or another company's server is billed. This is where the headline rates apply.

Inter-region transfer is data moving between two of the provider's own data centre regions — say from AWS Mumbai (ap-south-1) to Singapore (ap-southeast-1). Businesses trigger this constantly with cross-region replication, disaster recovery, and analytics pipelines.

Intra-region / inter-AZ transfer is the sneakiest one. Even within Mumbai, if your application server sits in one Availability Zone and your database in another, you pay a small per-GB fee on the traffic between them. At scale, "small" adds up fast.

Ingress — data coming into the cloud — is free on all three major providers. That asymmetry is deliberate. It's cheap and easy to move in, expensive to move out. Once your data and application logic live inside a provider's ecosystem, egress fees become a soft form of lock-in.

How much do egress fees actually cost in India in 2026?

Rates vary by provider, region, and volume tier, and they're quoted in USD, so your rupee cost floats with the exchange rate plus 18% GST on top. Here's a realistic comparison for workloads hosted in the Mumbai region, based on published list pricing for the first few TB per month. Treat these as planning figures, not billing guarantees.

Transfer type (Mumbai region)	AWS (approx)	Google Cloud (approx)	Azure (approx)
Internet egress (per GB, first 10 TB)	₹9.0	₹9.8	₹7.3
Free egress tier per month	100 GB	200 GB (general)	100 GB
Inter-region (Mumbai → Singapore, per GB)	₹7.5	₹8.0	₹6.6
Inter-AZ within region (per GB, each direction)	₹0.85	₹0.85	Free (same region VNet)*
Egress via CDN (per GB, India)	₹6.5 (CloudFront)	₹6.0 (Cloud CDN)	₹5.5 (Azure CDN)

*Azure's intra-region VNet rules have exceptions; availability-zone billing applies to some scenarios. Always confirm against the calculator for your architecture.

Notice two things. First, egress to the internet is not far off inter-region transfer — many teams assume "keeping it inside AWS" is free. It isn't. Second, a CDN is cheaper per GB than raw egress on every provider, which is why content-heavy sites should never serve directly from origin.

Why do Indian SMBs get hit harder than they expect?

A few patterns specific to how Indian businesses build things push egress bills up.

Split-region architectures. Plenty of teams still put their primary workload in Singapore because they set it up years ago before Mumbai and Hyderabad regions matured, then added an Indian region later for compliance. Now every sync between them is inter-region traffic.

Data residency rules driving replication. With India's data protection regime tightening, many businesses replicate data into an Indian region for compliance while keeping analytics elsewhere. That replication is billable. If you're navigating this, our breakdown of data residency rules in India for 2026 is worth reading before you design the topology.

Backups and DR to another region. Cross-region snapshot copies for disaster recovery generate steady egress that finance never sees coming.

Media-heavy apps. E-commerce product images, video KYC recordings, PDF invoices — Indian consumer apps push a lot of bytes to a lot of devices on patchy networks, which means retries and duplicate downloads.

Common Mistake: Teams enable cross-region automated backups for "safety" and forget them. I've seen a Pune SaaS firm copying 400 GB of RDS snapshots daily from Mumbai to Singapore — around ₹90,000/month in transfer alone — when a same-region backup plus a weekly cross-region copy would have met their actual RTO. Match your DR spend to your real recovery objective, not to a checkbox someone ticked in 2021.

Case study: how a Gurgaon logistics firm cut its transfer bill by 71%

A 40-person logistics and last-mile delivery company in Gurgaon runs a driver app, a customer tracking portal, and an internal dashboard. Their AWS bill was ₹3.6 lakh/month, of which data transfer was ₹1.15 lakh. Here's what we found and fixed over a six-week engagement.

The findings:

• The tracking portal served map tiles and vehicle images directly from an S3 bucket with public read — no CDN. That was ~2.5 TB/month of raw internet egress.
• Their primary database was in Singapore (legacy), while the app servers had been migrated to Mumbai. Every query crossed regions.
• Automated RDS snapshots were being copied to two other regions daily.
• Application and DB tier sat in different AZs with chatty, uncompressed API calls between them.

What we did:

1. Put CloudFront in front of S3. Cached map tiles and images at the edge. Egress dropped from ₹9/GB origin serving to ₹6.5/GB CDN, and cache hit ratio hit 84%, so most requests never touched origin at all. Saving: ~₹58,000/month.
2. Consolidated to a single region. Migrated the database into Mumbai (ap-south-1) alongside the app tier. Inter-region query traffic went to zero. This also improved latency for Indian users by roughly 40ms. Saving: ~₹22,000/month.
3. Right-sized DR. Replaced daily three-region snapshot copies with a same-region automated backup plus one weekly cross-region copy to Hyderabad. Saving: ~₹19,000/month.
4. Enabled gzip and reduced chatty calls. Compressed API payloads and batched several small calls. Inter-AZ traffic fell noticeably. Saving: ~₹8,000/month.

Total transfer cost went from ₹1.15 lakh to about ₹33,000/month — a 71% reduction, with better performance for users. The migration and re-architecture, including our fee, paid for itself in under three months. If you want that kind of review, our cloud migration and managed services team runs exactly this playbook.

How do I audit my own cloud data transfer costs?

You don't need a consultant to find the biggest leaks. Here's the process I use on day one of any engagement.

1. Open the cost breakdown by usage type. In AWS Cost Explorer, group by "Usage Type" and filter for anything containing DataTransfer or Out-Bytes. In GCP, use the billing report grouped by SKU and search "egress." In Azure, filter Cost Analysis by "Bandwidth" meter category.
2. Identify the top three transfer line items. Ninety percent of the time, three items account for most of the cost. Note whether each is internet egress, inter-region, or inter-AZ.
3. Map each line item to a workload. Which service, bucket, or database is generating it? Tag your resources by project so this is easy next time.
4. Check for a CDN gap. Is any static or media content served straight from origin? That's your fastest win.
5. Check your region topology. Are app, database, and cache all in the same region? Any cross-region link deserves scrutiny.
6. Review DR and backup replication. List every cross-region copy job and confirm it maps to a documented recovery requirement.
7. Model the fix. Estimate savings before you change anything, so you can prove ROI to whoever signs off.

Pro Tip: Set a billing alert specifically scoped to data transfer, not just total spend. Most teams alert on the total bill, which hides a transfer spike behind normal compute variation. A dedicated egress budget alert catches a runaway backup job or a viral content event within a day instead of at month-end.

What architectural choices keep transfer costs low?

Beyond the quick audit fixes, a few design principles keep egress structurally low over the long run.

Use a CDN for anything served repeatedly

Images, JavaScript bundles, videos, downloadable invoices, APK files — if the same bytes go to many users, cache them. CloudFront, Cloud CDN and Azure CDN all price India delivery below raw origin egress, and the caching means far fewer origin fetches.

Keep chatty services in one region and one AZ where sensible

There's a trade-off with high availability, but for many SMB workloads a well-designed single-region deployment with cross-region backup is plenty resilient. Don't pay AZ-crossing fees for a service that doesn't need multi-AZ.

Use private endpoints instead of public routes

Accessing S3, DynamoDB or other managed services over a VPC endpoint keeps traffic on the provider's private network and often avoids NAT gateway data processing charges — which are a hidden cousin of egress fees.

Compress everything

Gzip or Brotli on API responses and web assets can halve transferred bytes. It's a one-line config change on most stacks and it directly reduces the GB you're billed for.

Consider a sovereign or India-first cloud where it fits

For workloads with strict residency needs, keeping data and processing inside India avoids expensive cross-border transfer entirely. Our guide to the India sovereign cloud landscape for SMBs covers when this makes sense.

Should egress fees change my provider or migration decision?

Sometimes, yes. Egress is the single biggest reason a "cloud is cheaper" business case flips negative after migration. If your workload pushes huge volumes to the internet — a video platform, a large media library, a public API with heavy responses — the egress bill can dwarf your compute savings.

This is exactly why I insist on modelling transfer costs before any migration, not after. A proper cloud cost optimisation exercise treats egress as a first-class number in the projection. If the numbers are ugly, options include a CDN-heavy design, a provider with more generous egress terms, or in extreme cases a hybrid setup where high-egress serving stays on cheaper bandwidth.

Choosing between providers isn't only about egress rates, though. Support quality, managed service maturity, and your team's existing skills matter more for most SMBs. Our IT consulting practice helps weigh these against your actual workload rather than a generic price sheet.

Where eDarpan fits in

Most of the savings above come from experience, not magic — knowing where cloud providers hide the meter and having done enough migrations to spot the patterns fast. That's the work our cloud services team does daily for Indian SMBs.

If your monthly cloud bill has crept up and nobody can quite explain why, a focused cost audit usually pays for itself in the first month. We also help with the wider stack — custom software development that's architected to be cheap to run, mobile apps that don't hemorrhage bandwidth, and productivity licensing through Google Workspace or Microsoft 365. If you're setting up a new entity and need a virtual office address for GST and company registration, we handle that too. Reach out via our contact page for a no-obligation review.

Frequently asked questions about cloud data transfer costs in India

Is data transfer into AWS, GCP or Azure free in India?

Yes. Ingress (data coming into the cloud) is free on all three major providers, in the Indian regions and everywhere else. You're charged when data leaves the provider's network or moves between regions and, in some cases, between availability zones.

How much does AWS charge for egress from the Mumbai region?

As a planning figure, internet egress from ap-south-1 (Mumbai) runs around ₹9 per GB for the first several TB per month, after the first 100 GB free tier, plus 18% GST. Rates drop at higher volume tiers and are lower when serving through CloudFront. Always confirm against the AWS pricing calculator for your exact volume.

Does a CDN reduce cloud egress fees?

Yes, in two ways. CDN delivery is usually priced below raw origin egress per GB, and caching means most requests are served from the edge without ever fetching from your origin. For image and video-heavy Indian apps, a CDN commonly cuts effective transfer cost by 60–80%.

Are transfers between availability zones really billed?

On AWS and GCP, yes — inter-AZ traffic within a region is billed at roughly ₹0.85 per GB in each direction. It sounds trivial but adds up for chatty microservices or a database in a different AZ from the app tier. Azure's rules differ and many intra-region VNet flows are free, but confirm for your specific setup.

Will Indian data residency rules increase my transfer costs?

They can, because meeting residency often means replicating data into an Indian region while other systems live elsewhere, and that replication is billable inter-region traffic. The fix is usually to keep processing and storage in the same Indian region rather than splitting across borders. Our data residency guide explains the trade-offs.

Which cloud provider has the cheapest egress for Indian businesses?

On list pricing, Azure's Mumbai egress tends to be slightly lower per GB than AWS or GCP, but headline rates rarely decide the total bill. Free tiers, CDN pricing, committed-use discounts, and how well your architecture avoids transfer matter far more. Model your actual workload rather than picking on a per-GB number alone.

How do I stop surprise egress spikes on my bill?

Set a billing alert scoped specifically to data transfer usage types, tag resources so you can trace spikes to a workload, and review cross-region backup jobs quarterly. A runaway snapshot copy or a piece of viral content should trigger an alert within a day, not surface as a shock at month-end.

Final word on controlling cloud data transfer costs in India

Egress fees aren't a scam, but they are a trap for teams that don't watch them. The good news is that the fixes are well understood and mostly one-time: put a CDN in front of your content, keep chatty services in one region, right-size your DR, and compress your payloads. Do those four things and you've addressed the bulk of what inflates cloud data transfer costs India businesses quietly absorb every month.

Start with the audit. Open your cost breakdown, find the top three transfer line items, and map them to workloads. You'll almost certainly spot at least one easy win in the first hour. And if you'd rather have someone who's done this dozens of times run the review with you, that's exactly what our services team is here for — get in touch and we'll take a look at your bill.

Image credit: Innovate Maryland Emerging Technology Center by MDGovpics via flickr (BY 2.0), sourced through Openverse.