The “build vs. buy” decision for cloud operations is one of the most consequential — and least rigorously analysed — choices an Indian IT leader makes. Most comparisons stop at salaries. The real cost picture is two to three times more complex, and in the Indian context, attrition alone changes the entire calculus.
This post lays out a complete cost framework for the decision: what an in-house AWS team actually costs when you account for every line item, what MSP pricing actually buys you, and the scenarios where each model makes clear financial sense. The numbers are grounded in Indian market data, not US benchmarks.
How to use this post: Read through the full framework to understand the cost drivers, then jump to the decision matrix at the end to map your situation to the right model. The cost tables are designed to be adapted to your specific headcount and AWS spend.
Why Most Build-vs-Buy Analyses Get It Wrong
The typical framing of this decision goes: “Three cloud engineers at ₹15 LPA each costs ₹45 lakhs a year. An MSP would cost ₹30 lakhs. Therefore, we should hire.” This analysis fails for three reasons.
- It counts only salaries, not total employment cost. PF, gratuity, health insurance, payroll tax, IT equipment, software licences, office space, and HR overhead add 30–50% to base salary in India.
- It ignores attrition. Cloud engineers in India change jobs at rates well above the IT sector average. The cost of a single unfilled role — recruitment fees, ramp-up time, knowledge loss — often exceeds six months of salary.
- It conflates headcount with coverage. Three engineers covering Monday to Friday, 9am to 6pm is not the same as 24/7 operations with on-call escalation. The in-house model that truly replicates MSP coverage requires a larger team than most people plan for.
The honest comparison is not salary vs. retainer. It is the total cost of operating cloud infrastructure to a defined service level, via each model.
The Full Cost of an In-House AWS Team in India
Let us build a realistic in-house team for a mid-market Indian enterprise running meaningful workloads on AWS — say ₹15–30 lakhs per month in AWS spend. This is a company that needs reliable 24/7 operations, security management, and some capacity for architecture and optimisation work.
Minimum viable team composition
🏢 Roles Required
- 1× Cloud Architect / Lead (5–8 yrs)
- 2× Senior Cloud / DevOps Engineers (3–5 yrs)
- 1× Cloud Security Engineer (3–5 yrs)
- 1× Junior Cloud Engineer / L1 Support (1–3 yrs)
📋 Coverage Provided
- Architecture, migration, optimisation
- Day-to-day ops, IaC, automation
- IAM, GuardDuty, patching, compliance
- Monitoring, alerting, on-call rotation
Note on 24/7 coverage: Five engineers cannot maintain genuine 24/7 on-call coverage without significant burnout risk. In practice, most in-house teams cover business hours reliably and handle off-hours incidents on an ad-hoc basis. Replicating true round-the-clock operations with an in-house team requires at minimum 7–8 engineers when shifts, leave, and attrition buffers are factored in.
Salary benchmarks — India, 2026
The following ranges are based on Glassdoor data (March 2026) and publicly available salary surveys for cloud roles in India’s major tech hubs. Ranges reflect the 25th–75th percentile for Bangalore and Hyderabad, which is where most cloud talent is concentrated.
| Role | Experience | Salary Range (₹ LPA) | Midpoint Used |
|---|---|---|---|
| Cloud Architect / Lead | 5–8 yrs, AWS Pro cert | ₹22 – 40 LPA | ₹28 LPA |
| Senior Cloud / DevOps Engineer | 3–5 yrs, AWS Associate cert | ₹12 – 22 LPA | ₹16 LPA |
| Senior Cloud / DevOps Engineer #2 | 3–5 yrs, AWS Associate cert | ₹12 – 22 LPA | ₹16 LPA |
| Cloud Security Engineer | 3–5 yrs, AWS Security Specialty | ₹15 – 28 LPA | ₹20 LPA |
| Junior Cloud Engineer / L1 | 1–3 yrs, AWS CLF/SAA | ₹6 – 10 LPA | ₹8 LPA |
| Total Annual Salary (base) | ₹88 LPA | ||
True employment cost — beyond salary
Base salary is only part of what a headcount costs the organisation. The following table maps all additional cost categories that are typically budgeted separately but belong in the same comparison.
| Cost Category | Basis | Annual Estimate (₹) |
|---|---|---|
| Statutory and Payroll Costs | ||
| Employer PF contribution | 12% of basic (approx. 40% of CTC) | ₹4.2 L |
| Employee State Insurance / ESIC | 3.25% of gross (for eligible employees) | ₹1.1 L |
| Gratuity accrual | 4.81% of basic per year | ₹1.7 L |
| Payroll processing and compliance | Flat or per-head | ₹0.6 L |
| Benefits and Perks | ||
| Group health insurance | ₹15,000–25,000 per head per year | ₹1.0 L |
| Term life / personal accident cover | ₹5,000–8,000 per head per year | ₹0.35 L |
| Leave encashment, LTA, performance bonus | Typically 10–15% of base | ₹8.8 L |
| Infrastructure and Tooling | ||
| Laptops and workstations (amortised 3 yrs) | ₹80,000–1,20,000 per head | ₹1.7 L |
| Software licences (IDEs, monitoring tools, security tools) | Per-seat commercial licences | ₹3.5 L |
| SIEM / observability platform | Datadog, New Relic, or equivalent | ₹5.0 L |
| Security tooling (CSPM, vulnerability scanner) | Commercial tools not included in AWS billing | ₹3.0 L |
| Training and Certifications | ||
| AWS certification prep and exam fees | ₹20,000–35,000 per cert attempt | ₹1.5 L |
| Ongoing training budget | Industry benchmark: ₹30,000–50,000 per person | ₹2.0 L |
| Recruitment and Overhead | ||
| Initial recruitment (placement agency or job boards) | 8–12% of CTC for agency; amortised over 2 yrs | ₹5.3 L |
| HR time and onboarding | Estimated 40–60 hours per hire × internal HR cost | ₹1.2 L |
| Office space allocation (if applicable) | ₹8,000–15,000 per seat per month in metro | ₹7.2 L |
| Total Additional Costs (annual) | ₹48.15 L | |
Total in-house cost before attrition: ₹88 L (salaries) + ₹48 L (additional costs) = approximately ₹1.36 crore per year for a five-person team. This is the baseline — before accounting for the most disruptive cost driver in the Indian market.
The Attrition Multiplier — India’s Hidden Cost
India’s IT sector attrition rate is approximately 25% across the industry, with cloud and cybersecurity roles experiencing even higher churn due to intense competition for niche skills. For a five-person cloud team, this means statistically losing at least one to two engineers every year.
The cost of each departure compounds in three ways:
1. Direct replacement cost
- Recruitment agency fee: 8–12% of annual CTC for a specialised cloud role. For a ₹16 LPA engineer, that is ₹1.3–1.9 lakhs per hire.
- Time-to-hire for senior cloud roles in India: 45–90 days on average, per recruiter data.
- During vacancy, remaining team members cover the workload — increasing burnout risk and triggering further attrition.
2. Ramp-up and productivity loss
- A new cloud engineer reaching full productivity in a complex environment typically takes 60–90 days.
- During this period, the team operates at reduced capacity — incidents take longer to resolve, optimisation work stalls, and the architect or lead spends disproportionate time on knowledge transfer.
- For a senior engineer, the combined cost of vacancy and ramp-up loss is estimated at 40–60% of their annual salary per departure event.
3. Knowledge and institutional loss
- Cloud environments accumulate undocumented knowledge: why certain architectural decisions were made, quirks in the production setup, institutional understanding of the company’s compliance posture.
- When this knowledge walks out the door, incidents become harder to diagnose and architecture decisions get repeated. This cost is real but rarely quantified.
The India-Specific Compounding Effect
India’s cloud talent market has an additional dynamic: professionals with AWS certifications and 3+ years of experience have multiple competing offers at any time. A certified AWS Solutions Architect Professional with security experience is in a seller’s market. Retention strategies — counter-offers, ESOP, flexibility — add their own cost. And attrition in cloud teams tends to cluster: when a lead leaves, junior engineers often follow within 6 months, creating team-level disruption rather than individual turnover.
Quantifying attrition cost for a five-person team
| Attrition Scenario | Departures/Year | Cost per Departure | Annual Attrition Cost |
|---|---|---|---|
| Conservative (20% rate) | 1 person | ₹6–9 L | ₹6–9 L |
| Typical IT sector (25% rate) | 1–2 persons | ₹6–9 L each | ₹9–18 L |
| Cloud / niche roles (30%+ rate) | 2 persons | ₹9–12 L each | ₹18–24 L |
Adding a conservative attrition cost of ₹12 lakhs per year brings the realistic annual cost of the in-house model to approximately ₹1.48 crore — and that is a good year. A year with a lead departure and a security engineer departure can push the total well above ₹1.6 crore.
What AWS MSP Pricing Actually Covers
MSP pricing in India is structured in two primary models, often combined:
Model 1: Percentage of AWS spend
The most common structure for managed services is a retainer calculated as a percentage of monthly AWS spend — typically 8–15% for Indian MSPs, with the range driven by the scope of services included (monitoring only vs. full ops + security + FinOps).
For a company spending ₹20 lakhs per month on AWS:
- At 10%: ₹2 lakhs/month = ₹24 lakhs/year in management fees
- At 12%: ₹2.4 lakhs/month = ₹28.8 lakhs/year
- At 15%: ₹3 lakhs/month = ₹36 lakhs/year
Model 2: Fixed monthly retainer + scope-based billing
Some MSPs, particularly for smaller AWS footprints, charge a fixed monthly retainer for a defined scope — for example, ₹1.5–2.5 lakhs per month for 24/7 monitoring, incident response, patching, and basic FinOps — with additional charges for project work (migrations, major architecture changes, security audits).
What the retainer typically includes
Usually Included
- 24/7 infrastructure monitoring and alerting
- Incident response and resolution (P1–P4)
- OS and security patching
- IAM governance and access reviews
- Monthly cost and optimisation report
- GuardDuty / Security Hub management
- RI and Savings Plan recommendations
- Backup monitoring and validation
Typically Add-On
- Major architecture changes or migrations
- New environment buildout
- Penetration testing / security audits
- Compliance audit preparation (SOC 2, ISO)
- Custom application deployment pipelines
- Data analytics and BI workloads
- Training for your internal team
The MSP cost advantage is not just the retainer fee. An MSP’s tooling — SIEM platform, CSPM, vulnerability scanner, observability stack — is shared across their customer base. The per-customer cost of this tooling in an MSP model is a fraction of what it costs to procure and maintain the same stack independently. For Indian enterprises, this tooling cost differential alone can justify the MSP premium.
Side-by-Side: A Realistic Cost Model for an Indian Mid-Market Enterprise
The following comparison uses a consistent scenario: a Bangalore-based enterprise spending ₹20 lakhs per month on AWS, needing reliable 24/7 operations, security, and FinOps management.
| Cost Category | In-House Team (Annual ₹) | AWS MSP (Annual ₹) |
|---|---|---|
| People Costs | ||
| Base salaries (5 FTEs) | ₹88,00,000 | — |
| PF, gratuity, statutory contributions | ₹7,00,000 | — |
| Benefits (insurance, bonus, LTA) | ₹10,15,000 | — |
| Recruitment (initial + ongoing) | ₹6,50,000 | — |
| MSP retainer fee (12% of ₹20L/mo) | — | ₹28,80,000 |
| Tooling and Infrastructure | ||
| Laptops and workstations (amortised) | ₹1,70,000 | — |
| Software licences | ₹3,50,000 | — |
| SIEM / observability platform | ₹5,00,000 | Included |
| CSPM / security tooling | ₹3,00,000 | Included |
| Training and Certifications | ||
| AWS certification and training | ₹3,50,000 | — |
| Overhead | ||
| Office space allocation | ₹7,20,000 | — |
| HR, payroll processing, admin | ₹1,80,000 | — |
| Attrition (conservative estimate) | ||
| Replacement recruitment + ramp-up loss | ₹12,00,000 | — |
| Total Annual Cost | ₹1,49,35,000 | ₹28,80,000 |
Why the gap is so large: The MSP model does not carry people costs, tooling procurement, attrition risk, or office overhead. A ₹28–36 lakh MSP retainer for a ₹20 lakh/month AWS spend sounds significant until it is placed next to ₹1.5 crore in total in-house operating cost. The real question is never retainer vs. salary — it is total cost of operations to a defined service level.
The INR/USD Factor
AWS pricing is denominated in USD, which means Indian enterprises absorb currency risk when their cloud spend is not hedged. An MSP operating on a fixed INR retainer partially insulates you from exchange rate volatility on the management cost side. While the underlying AWS charges still follow the USD rate, at least the operational layer is priced in local currency with predictable monthly billing.
What Money Does Not Capture — Capability and Risk Gaps
The financial comparison above strongly favours the MSP model for most Indian mid-market enterprises. But cost is not the only dimension. There are capability and risk factors that do not appear in any spreadsheet.
Arguments for in-house that go beyond cost
- Deep product knowledge: An in-house engineer who understands the business domain, the codebase, and the infrastructure together can make architectural decisions that an MSP engineer cannot. This matters most for product companies where cloud infrastructure is tightly coupled to application behaviour.
- Speed of response on complex issues: For incidents that require both application context and infrastructure knowledge simultaneously, an in-house team is faster because they hold both.
- Regulatory sensitivity: In highly regulated environments (government, defence, certain BFSI) where data exposure to any third party creates legal risk, in-house may be mandated regardless of cost.
- Strategic control: If cloud architecture is a core competitive differentiator — as it is for large-scale SaaS companies — ceding that to an MSP creates dependency that may be difficult to unwind.
Arguments for MSP that go beyond cost
- Breadth of expertise: A good MSP’s team has seen hundreds of environments. The depth of exposure to failure modes, edge cases, and optimisation opportunities that comes from managing diverse workloads at scale is genuinely hard to replicate in a five-person in-house team.
- No single points of failure: If your lead cloud architect leaves, your MSP’s service level does not change. The institutional knowledge risk that devastates in-house teams does not apply.
- Tooling ecosystem: MSPs have enterprise agreements with monitoring, security, and automation vendors that smaller in-house teams cannot access at equivalent cost.
- Regulatory currency: A good MSP actively tracks AWS service updates, new compliance requirements (DPDP Act, CERT-In advisories), and security advisories as part of their core business. Staying current is an overhead that in-house teams often underinvest in.
Three Scenarios and the Model That Fits Each
Scenario A
The D2C / E-commerce Scale-Up
₹8–25L/mo AWS spend, growing fast, engineering team focused on product, no dedicated cloud ops person, experiencing increasing incidents and cost unpredictability.
→ MSP model
Scenario B
The Large Enterprise / GCC
₹50L+/mo AWS spend, cloud is core to competitive differentiation, existing in-house team of 8–15, complex multi-account architecture, regulatory requirements.
→ In-house + augmentation
Scenario C
The Mid-Market Manufacturer / BFSI
₹15–40L/mo AWS spend, IT team focused on business applications, cloud is infrastructure not product, compliance pressure increasing, limited cloud expertise in-house.
→ MSP model
The hybrid model — an internal cloud lead or architect who owns strategy and vendor relationship, supported by an MSP for day-to-day operations and 24/7 coverage — is increasingly the most cost-efficient structure for mid-market enterprises. It captures the domain knowledge advantage of in-house while offloading the operational overhead and tooling cost to the MSP.
The Decision Matrix
Map your situation against the following factors to identify the model that is likely to deliver the best outcome for your organisation.
| Factor | In-House Favoured | MSP Favoured |
|---|---|---|
| AWS monthly spend | ₹50L+ / month | Below ₹40L / month |
| Cloud’s role in the business | Core product differentiator | Infrastructure / utility |
| Existing team depth | 8+ cloud engineers already in place | 0–3 cloud-focused engineers |
| Coverage requirement | Business hours ops acceptable | 24/7 SLA-backed coverage required |
| Security and compliance | Mandated in-house (defence, govt) | DPDP, RBI, CERT-In compliance needed |
| Tolerance for attrition risk | High — strong retention programme | Low — team continuity is critical |
| FinOps maturity needed | Team has or can develop capability | Need structured RI management now |
| Speed to operational readiness | Can afford 3–6 month ramp-up | Need coverage in 30–60 days |
The Real Question
The build-vs-buy decision for cloud operations is ultimately not a cost decision — it is a strategic one about where your organisation’s engineering attention should be focused. For most Indian mid-market enterprises, the in-house model costs two to four times more than an MSP arrangement when all costs are accounted for, and it comes with structural risks (attrition, knowledge concentration, tooling gaps) that the MSP model eliminates.
The enterprises for whom in-house investment is clearly justified are those where cloud architecture is genuinely a source of competitive advantage — where the depth of domain-specific cloud expertise translates directly into product or business outcomes that an external partner cannot replicate.
For everyone else, the question is not “should we use an MSP?” but “what should we retain in-house, and what should the MSP own?” That framing leads to the hybrid model — and to a much more honest conversation about where your engineering talent is best spent.
AWS Managed ServicesIn-House vs MSPCloud Cost IndiaCloud Engineer Salary IndiaIT Attrition IndiaFinOpsBuild vs Buy CloudAWS IndiaCloud StrategyCTO Decision Framework
