Photo: Unsplash — free to use, no attribution required
Moving workloads off on-premise servers to the cloud is one of the most consequential infrastructure decisions an Indian business makes in 2026. The promise — lower capex, elastic scaling, better uptime — is real. But so are the surprises that arrive on the first month's bill after a migration that was planned on assumptions borrowed from US pricing guides. This guide is built specifically around Indian market realities: Mumbai region pricing, rupee cost estimates, RBI and DPDPA compliance obligations, and the honest hidden costs that catch companies off guard.
AWS Mumbai Region (ap-south-1): What Indian Businesses Actually Pay
AWS has operated its Mumbai region (ap-south-1) since 2016 and it remains the primary choice for Indian cloud deployments. One number every Indian cloud buyer needs to understand upfront: the Mumbai region carries a 15–20% pricing premium over AWS US-East-1. This reflects the cost of local infrastructure, lower utilisation density, and data sovereignty overhead. It is not a negotiable surcharge — it is baked into published pricing.
For the most commonly used compute and database instance types, on-demand pricing in ap-south-1 works out as follows:
- EC2 t3.medium (2 vCPU, 4 GB RAM): $0.0464/hour — approximately ₹3.85/hour, or ₹2,780/month running continuously
- EC2 m5.large (2 vCPU, 8 GB RAM): $0.096/hour — approximately ₹8.00/hour, or ₹5,760/month
- EC2 c5.xlarge (4 vCPU, 8 GB RAM): $0.17/hour — approximately ₹14.10/hour
- RDS db.t3.medium MySQL (2 vCPU, 4 GB RAM): $0.068/hour — approximately ₹5.65/hour, or ₹4,070/month
- RDS db.r5.xlarge (4 vCPU, 32 GB RAM): $0.48/hour — approximately ₹39.90/hour
- S3 Standard storage: $0.025/GB/month — approximately ₹2.08/GB/month
These are on-demand rates. Moving to 1-year reserved instances immediately cuts compute costs by 30–38%. A t3.medium on a 1-year reserved term in Mumbai drops to approximately $0.0295/hour (₹2.45/hour). For workloads that run continuously, reserved instances are the correct pricing model, not on-demand.
AWS Marketplace India lists services from Indian ISVs and global SaaS vendors with local billing. For managed migration services, the major Indian IT firms — TCS Cloud Division, Infosys Cobalt, and Wipro FullStride Cloud — provide end-to-end migration engagements typically priced as a percentage of the first year's estimated AWS spend (usually 20–35% of Year 1 cloud spend as professional services fees). Independent cloud consultants and boutique firms offer the same technical capability at 40–60% lower professional services cost, which matters significantly for mid-market Indian companies where the migration project budget is itself under pressure.
Azure in India: Central India (Pune) and South India (Chennai)
Microsoft Azure operates two primary Indian regions — Central India (Pune) and South India (Chennai) — plus a West India region (Mumbai) that is available for replication but not open to new deployments. For most Azure India deployments, Central India (Pune) is the recommended primary region, with South India (Chennai) as the paired disaster recovery region.
What genuinely differentiates Azure for Indian enterprises is compliance pedigree in the regulated sector. Microsoft has received approval from the Reserve Bank of India for its Azure Government data centre infrastructure — a distinction that matters enormously for BFSI companies (banks, NBFCs, insurance firms) that must comply with RBI's 2023 Master Direction on IT Governance and IRDAI's cloud guidelines. When your auditors ask for a cloud compliance attestation that satisfies RBI examiners, Azure is currently the most straightforward answer.
Azure compute pricing in India is broadly comparable to AWS, with a few important differences:
- Azure B2s VM (2 vCPU, 4 GB RAM — closest equivalent to t3.medium): approximately $0.052/hour in Central India (₹4.32/hour)
- Azure D2s v3 (2 vCPU, 8 GB RAM): approximately $0.096/hour (₹8.00/hour)
- Azure SQL Database (General Purpose, 2 vCores): approximately $0.34/hour (₹28.25/hour)
- Azure Blob Storage (Hot tier): $0.0208/GB/month (₹1.73/GB/month)
Azure Hybrid Benefit is the single biggest cost variable for Indian enterprises. If your company runs existing Windows Server or SQL Server licences under Software Assurance, you can apply those licences to Azure VMs and avoid paying the Windows OS licence component. On a D4s v3 Windows VM, Azure Hybrid Benefit reduces the VM cost by 40–80% depending on the licence tier. For an Indian company migrating 20 Windows servers to Azure, this benefit alone can save ₹6–15 lakh per year in licence costs compared to running equivalent AWS Windows instances. Any Indian enterprise with a Microsoft volume licensing agreement should run an Azure Hybrid Benefit calculator before finalising a cloud vendor decision.
Google Cloud Mumbai (asia-south1): The Analytics-First Option
Google Cloud's Mumbai region (asia-south1) is the newest of the three major cloud presences in India and carries a smaller market share than AWS or Azure among Indian enterprise customers. That market share gap is primarily a sales and partner ecosystem issue, not a technical one — the infrastructure is solid and latency to most Indian metros is competitive.
Where Google Cloud genuinely differentiates for Indian businesses is in analytics and data warehouse workloads. BigQuery, Google's serverless data warehouse, has no real equivalent in the AWS or Azure managed service portfolio at the same price-to-performance ratio. For Indian companies running large data analytics pipelines — particularly fintech companies doing transaction analysis, healthcare companies processing patient records at scale, or e-commerce platforms doing behavioural analytics — GCP's asia-south1 region with BigQuery is worth a detailed evaluation.
GCP compute pricing in Mumbai:
- e2-medium (2 vCPU, 4 GB RAM): approximately $0.042/hour (₹3.49/hour) — slightly below AWS t3.medium on-demand
- n2-standard-2 (2 vCPU, 8 GB RAM): approximately $0.097/hour (₹8.06/hour)
- Cloud SQL db-n1-standard-2 (2 vCPU, 7.5 GB RAM): approximately $0.15/hour (₹12.45/hour)
- BigQuery storage: $0.023/GB/month (₹1.91/GB/month); query processing $5/TB
GCP's committed use discounts (equivalent to AWS reserved instances) offer 28% savings on 1-year and 46% on 3-year commitments in Mumbai. Google Workspace customers who already pay for Google productivity tools will find GCP integration seamless and may benefit from negotiated bundle pricing.
Migration Types and Realistic Indian Cost Estimates
Cloud migration is not a single activity — it is a spectrum of technical approaches, each with meaningfully different cost profiles. For Indian companies migrating from on-premise to cloud, the three primary approaches and their realistic 2026 cost estimates are:
Lift-and-Shift (Rehosting)
Moving virtual machines or physical servers to cloud as-is, with minimal code changes. Fastest to execute, lowest migration cost, but does not capture the full efficiency benefits of cloud-native architecture.
- Scope: 20-server on-premise setup
- Migration project cost (professional services): ₹5–15 lakh one-time
- Timeline: 6–16 weeks
- Best for: Companies with immediate data centre exit deadlines, legacy applications with unclear code ownership, or businesses where the primary goal is eliminating hardware refresh capex
Re-platforming (Lift-and-Reshape)
Moving workloads to cloud managed services — replacing self-managed MySQL with RDS, replacing self-managed Tomcat with Elastic Beanstalk or Azure App Service, replacing on-premise file servers with S3 or Azure Blob. This requires moderate application changes but significantly improves operational efficiency and cost.
- Migration project cost: ₹15–40 lakh
- Timeline: 3–6 months
- Best for: Companies willing to invest in the migration project for better long-term cloud economics and reduced ongoing management overhead
Re-architecting to Microservices or Serverless
Breaking monolithic applications into containerised microservices (running on ECS, EKS, AKS, or GKE) or serverless functions (AWS Lambda, Azure Functions, GCP Cloud Run). Highest migration investment, but delivers the best long-term scalability and cost-efficiency for high-traffic applications.
- Migration project cost: ₹40 lakh to ₹2,00,000+ (for large enterprise systems)
- Timeline: 6–18 months
- Best for: Product companies and SaaS platforms that need to scale unpredictably, or companies building new features concurrently with migration
India-Specific Compliance: RBI, IRDAI, and DPDPA 2023
Compliance obligations in India add both cost and architectural constraints to cloud migration projects that do not appear in generic cloud migration guides written for the US market.
RBI Cloud Guidelines (Banks and NBFCs): The Reserve Bank of India's 2023 Master Direction on IT Governance, Risk, Controls and Assurance Practices requires that banks and NBFCs store regulated customer data within India. All three major cloud providers have Mumbai-region data centres that satisfy the geographic requirement. However, RBI guidelines also mandate a right-to-audit clause in cloud contracts, detailed incident response SLAs, and board-approved cloud risk assessments. Building these into a cloud migration programme adds 4–8 weeks to the project timeline and typically requires dedicated compliance advisory work (budget ₹5–10 lakh for regulatory advisory if your legal team lacks cloud compliance expertise).
IRDAI Cloud Guidelines (Insurance Sector): The Insurance Regulatory and Development Authority of India has issued its own IT framework for insurance companies, which similarly requires data localisation and pre-migration regulatory intimation. Insurance companies planning a cloud migration should factor in a 60–90 day regulatory notice period before the cutover date.
DPDPA 2023 (All Sectors): The Digital Personal Data Protection Act 2023 introduces data residency obligations for personal data of Indian citizens. While the detailed rules are still being operationalised, the practical implication for cloud migrations is clear: workloads handling Indian personal data should run in Indian cloud regions by default. This effectively eliminates cost-optimisation strategies that involve routing Indian user data through cheaper Singapore or US-East regions — an approach some Indian startups used before DPDPA came into force.
For Technopark and Infopark companies in Trivandrum and Kochi processing data for international clients, DPDPA compliance adds a data classification layer to cloud architecture — you need to know precisely which data belongs to Indian citizens and ensure it stays in Indian regions. AWS Mumbai, Azure Central India, and GCP Mumbai all support this requirement natively.
TCO Analysis: On-Premise vs Cloud for Indian Businesses
The most honest way to evaluate cloud migration economics is a total cost of ownership (TCO) comparison that includes all cost categories — not just the monthly cloud bill.
On-Premise Server (Dell PowerEdge R750 Example)
- Hardware capex: ₹8–12 lakh for a mid-range server with 2 processors and 128 GB RAM
- Useful life: 3–5 years (after which hardware refresh is needed)
- Annual maintenance contract (AMC): ₹60,000–1,20,000/year
- Power and cooling: ₹40,000–80,000/year depending on data centre or office hosting
- Colocation or data centre rack space: ₹1,20,000–3,00,000/year in Tier-3 Indian data centres
- IT staff overhead: Shared cost of 1 sysadmin managing 10–20 servers — approximately ₹1–2 lakh/year per server when prorated
- Effective annualised cost (3-year horizon): ₹4–6 lakh/year per server including amortised capex
AWS Mumbai Equivalent (3-Year Reserved Instance)
- RDS db.r5.xlarge (4 vCPU, 32 GB RAM) — 3-year reserved, no upfront: approximately ₹1.8–2.5 lakh/year
- S3 storage (500 GB): approximately ₹12,000/year
- Data transfer and egress: ₹30,000–80,000/year depending on traffic
- AWS Business Support (10% of usage, min $100/month): ₹60,000–1,20,000/year
- Effective annual cost: ₹2.1–3.8 lakh/year
The cloud advantage at the infrastructure layer is real — roughly 30–50% lower annualised cost for compute and database workloads when using reserved instances. However, this saving narrows significantly once you add professional services for migration, staff retraining, and the one-time cost of addressing application compatibility issues. Most Indian mid-market companies reach positive ROI on their cloud migration investment within 18–30 months.
The Hidden Costs That Inflate Indian Cloud Bills
Every Indian cloud migration project that runs over budget does so because of costs that were either unknown or optimistically estimated at planning time. Here are the four categories that deserve their own budget line:
Data Egress Charges
AWS Mumbai charges $0.1086/GB for data transferred out of the region beyond the first 10 TB per month. A business with active reporting pipelines, external API consumers, or on-premise systems that pull data from cloud can easily accumulate 5–20 TB/month in egress. At 15 TB/month, the egress cost above the 10 TB threshold is approximately ₹41,500/month — an ongoing cost that most migration assessments do not model because on-premise networking costs are invisible (the bandwidth is already paid for).
Support Plan Costs
AWS Business Support is charged at 10% of monthly usage with a $100 minimum per month. At ₹5 lakh/month of AWS spend, Business Support adds ₹50,000/month. Azure and GCP have comparable paid support tiers. These support costs are mandatory for any business that needs human response to critical incidents — running production workloads on Developer Support is not a viable option.
Staff Retraining and Certification
Moving to cloud requires your IT team to genuinely upskill. An AWS Certified Solutions Architect — Associate course on Udemy India costs ₹15,000–40,000 per person (depending on whether the company pays for instructor-led training or self-paced courses). The AWS exam fee itself is $150 (approximately ₹12,500). A team of 3 engineers getting cloud-certified costs ₹80,000–1,50,000 in direct training costs, plus 3–6 months of reduced productivity as they move up the learning curve. Azure AZ-104 and GCP Professional Cloud Architect certifications have similar cost profiles.
Licensing Compliance
Oracle Database licences, certain Microsoft products, and other enterprise software have complex cloud licensing rules. Running Oracle Database on AWS EC2 requires an Oracle Database licence that is counted per physical core — not per vCPU — which can make Oracle licensing on cloud 2–4x more expensive than on-premise. Many Indian companies discover this only after migration, when Oracle's licence audit team arrives. This is one area where a detailed software asset management review before migration is essential, not optional. See also the CRM software guide for Indian businesses for how cloud-native SaaS alternatives can sometimes replace legacy licensed software entirely.
Kerala IT Companies and Cloud Migration in 2026
Trivandrum's Technopark and Kochi's Infopark host over 600 IT companies collectively, ranging from global product companies with 5,000+ employees to boutique software consultancies with 10–50 developers. The cloud migration patterns in Kerala's IT ecosystem reflect the national trends but with some local texture.
AWS has a registered partner network in Kerala, including several Technopark-based managed service providers who are AWS Advanced Consulting Partners. Azure's partner presence in Kochi is anchored by system integrators that grew out of Microsoft's long-standing enterprise relationships in Kerala's banking and government sectors. GCP's partner presence in Kerala is thinner but growing, primarily through analytics-focused boutiques.
For Kerala-based software product companies (ISVs) building for international markets, the US-East-1 versus Mumbai region decision is not straightforward. If 80% of your users are in North America, running your production workload in Mumbai adds 180–220ms of latency for your primary user base. Many Kerala ISVs run their production environment in US-East-1 or EU regions and use Mumbai purely for Indian market deployments or DPDPA-compliant data residency. This multi-region architecture adds cost complexity but is the correct long-term design for globally distributed SaaS companies.
For Kerala-focused businesses — retailers, hospitals, educational institutions, NBFCs, and logistics companies — AWS Mumbai or Azure Central India are the natural primary regions. The combination of low latency to Kerala users and Indian compliance credentials makes the choice straightforward.
Cloud migration costs are closely connected to the broader digital infrastructure decisions an Indian business makes. If you are also evaluating website development costs in India or mobile app development costs, the cloud hosting decision directly affects your ongoing operational expenses for both — choosing the wrong instance type or region at launch creates a cost overhang that compounds for years.
"The most expensive cloud migrations I have seen in India are not the ones with the highest AWS bills — they are the ones that were planned on spreadsheets built from US pricing guides. Mumbai region costs more than Virginia. Egress costs real money. Support plans are not optional when you run production workloads. And compliance work for RBI or DPDPA cannot be compressed out of the schedule. Plan for Indian realities from day one, not as a surprise in month three."
Frequently Asked Questions: Cloud Migration Cost in India 2026
What is the realistic total cost of migrating a 10-server on-premise setup to AWS Mumbai for an Indian mid-size company?
For a 10-server on-premise setup migrating to AWS Mumbai (ap-south-1) via lift-and-shift, a realistic total first-year budget falls between ₹8–18 lakh. This includes AWS EC2 compute (roughly ₹3–5 lakh/year for a mix of t3.medium and m5.large instances on 1-year reserved pricing), RDS managed database costs (₹1.5–3 lakh/year), S3 storage, data transfer, and migration professional services (₹2–5 lakh one-time). Reserved instances for a 1-year term save roughly 30–40% over on-demand Mumbai pricing. Year two onwards, costs typically drop 20–30% once the migration overhead is absorbed and reserved instances are fully in place.
Which cloud provider — AWS, Azure, or Google Cloud — is best for Indian businesses, and why?
The answer depends on your workload type and existing stack. AWS is the default for most Indian companies — largest partner ecosystem, most mature Mumbai region infrastructure, widest managed service catalogue. Azure is the better choice for enterprises with existing Windows Server or SQL Server licences: Azure Hybrid Benefit can save 40–80% on Windows VM costs, making Azure significantly cheaper on a total-cost basis. Azure also holds RBI data centre approvals, making it the cleaner compliance path for BFSI companies. Google Cloud's Mumbai region (asia-south1) has lower market share in India but is genuinely superior for large analytics workloads — BigQuery for high-volume data processing is a compelling differentiator. Evaluate GCP seriously if analytics is your primary cloud use case.
Does RBI's cloud framework affect which cloud provider an Indian BFSI company can choose?
Yes, materially. RBI's 2023 Master Direction on IT Governance requires banks and NBFCs to store regulated data in India, maintain a right-to-audit clause with cloud providers, and submit board-approved cloud risk assessments before migration. All three major providers technically satisfy the geographic data residency requirement. However, Microsoft has specific RBI-approved Government data centre infrastructure in India, giving Azure a compliance documentation advantage for BFSI audits. IRDAI has parallel requirements for insurance companies, including a 60–90 day regulatory notice period before cloud cutover. Budget 2–4 months for regulatory due diligence before the migration project start date if your organisation is RBI- or IRDAI-regulated.
How long does a typical cloud migration project take for an Indian SME?
For a business with 5–20 servers, a lift-and-shift migration typically runs 6–16 weeks from kick-off to production cutover. Re-platforming to managed cloud services takes 3–6 months. Full re-architecting to microservices or serverless is a 6–18 month programme. The most consistent cause of delays in Indian SME migrations is insufficient discovery work upfront — undocumented application dependencies, unlicensed software that needs remediation before cloud deployment, and database schema issues that only surface during migration testing. Budget 3–4 weeks for a thorough discovery and assessment phase before committing to a cutover date.
What hidden costs do Indian companies most often overlook when planning cloud migration?
Four categories catch Indian companies most often: (1) Data egress — AWS Mumbai charges $0.1086/GB beyond 10 TB/month; active reporting or API workloads can accumulate ₹40,000–1,50,000/month in egress that on-premise networking never surfaced. (2) Support plan costs — AWS Business Support at 10% of usage with a $100/month minimum adds material cost at any reasonable spend level. (3) Staff retraining — AWS Certified Solutions Architect courses on Udemy India cost ₹15,000–40,000 per person; the exam fee adds another ₹12,500; and the productivity dip during the learning period is an indirect cost that belongs in the business case. (4) Software licensing compliance — Oracle Database and certain Microsoft products have cloud-specific licence rules that can make cloud deployments 2–4x more expensive than on-premise if not identified during the pre-migration assessment.
