Netflix, Mercedes-Benz, ExxonMobil, and Nomura are among the most prominent AWS enterprise customers, and what they share is more instructive than their size. Across every large-scale migration, three patterns repeat: radical standardization before moving workloads, a proof-of-concept built around a real business workflow, and compliance automation treated as a first-order requirement. As of late 2024, approximately 2.38 million organizations use AWS infrastructure, spanning from early-stage startups to Fortune 500 companies. If you’re planning a migration, these are the patterns worth copying.
Three things to do before your next planning meeting:
- Prioritize workloads by business impact, not technical convenience
- Run a targeted PoC on one high-value workflow before committing to full scope
- Confirm compliance SLAs and regulatory lead times with your legal team now, not at cutover
Table of Contents
- What do AWS’s biggest clients actually run on the platform?
- What patterns do large AWS clients repeat across migrations?
- How should you plan a safe, efficient migration to AWS?
- When does it make sense to bring in an AWS migration partner?
- What measurable outcomes do enterprise AWS migrations actually produce?
- Key Takeaways
- The mistake most enterprises make — and what actually fixes it
- What IT-Magic’s AWS migration service delivers for your team
- Sources and further reading
- FAQ
What do AWS’s biggest clients actually run on the platform?
The most useful way to study big companies using AWS isn’t to admire their scale — it’s to map their workload types and migration strategies to your own situation.
| Organization | Workload Type | Primary Goal | Migration Strategy | Governance Approach | Measurable Impact |
|---|---|---|---|---|---|
| Netflix | Global streaming delivery | High availability at scale | Replatform / refactor | Cost governance, availability SLAs | 269M+ paid subscribers supported |
| Mercedes-Benz | SAP / mainframe consolidation | Simplification + AI enablement | Replatform (RISE with SAP) | Radical standardization, centralized governance | Reduced application count, cost savings |
| ExxonMobil | Collaboration platform (Digital Project Home) | Engineering productivity | PoC-first replatform | AWS Professional Services oversight | Up to 30,000 engineering hours saved per capital project |
| Nomura Group | Compliance automation, generative AI | Accuracy + processing speed | Refactor (AI-native) | Automated compliance agents | Reduced processing times for high-volume tasks |
| Samsung | User identity database | Cost reduction, latency | Replatform to Aurora | Phased migration, query validation | lower monthly costs vs. Oracle |
Netflix’s infrastructure story is worth a closer look for any team running high-availability services. The company supports over 269 million paid subscribers and is estimated to spend hundreds of millions annually on AWS, making it one of the platform’s largest spenders. For a deeper look at how Netflix structures its cloud infrastructure, the Netflix AWS breakdown covers the architecture specifics.
Mercedes-Benz’s SAP migration is the clearest enterprise example of standardization as a migration prerequisite. Before moving to AWS via RISE with SAP, the company eliminated redundant systems and centralized governance — reducing the application portfolio significantly. That groundwork is what made the migration manageable at global scale.
Nomura’s approach is the model for regulated financial services. The firm partnered with the AWS Generative AI Innovation Center to build compliance-checking agents directly into its infrastructure, handling high-volume workloads with automated accuracy checks. That’s not a post-migration add-on — it was designed in from day one.
What patterns do large AWS clients repeat across migrations?
The most reliable migration patterns from AWS client success stories aren’t proprietary. They’re repeatable, and they show up across industries.
Work backwards from a business workflow. ExxonMobil didn’t start with infrastructure. They started with a specific problem — fragmented collaboration across capital projects — and built a PoC around that workflow. The PoC delivered on time and on budget, which gave leadership the confidence to fund a broader rollout. That sequence matters: business problem first, platform second.
Standardize before you migrate. Mercedes-Benz’s lesson is that moving complexity to the cloud just makes it faster and more expensive. Eliminating redundant applications and centralizing governance before migration is what unlocks AI/ML capabilities afterward. BMW reinforced this: by refactoring for Amazon EKS, the company cut distributed AI model training from hours to minutes across its Connected AI Platform.
Automate compliance from the start. For Nomura, compliance automation wasn’t a phase-two project. It was the architecture. Any regulated enterprise — banking, healthcare, energy — should treat audit trails and automated checks as migration requirements, not enhancements.
Stage cutovers to minimize downtime. Nubank’s payments migration required advance coordination with Brazil’s central bank, adding necessary lead time to the project plan. Regulated workloads need regulatory calendars built into the migration schedule, not bolted on at the end.
Pro Tip: The most common anti-pattern in enterprise migrations is lift-and-shift without simplification. Moving a bloated application portfolio to AWS doesn’t reduce complexity — it preserves it at cloud speed and cloud cost. Rationalize first, migrate second.
How should you plan a safe, efficient migration to AWS?
The sequence that works at enterprise scale: Assess and rationalize → PoC a high-value workflow → Pilot → Scale. Every step below maps to that spine.
- Inventory and dependency mapping. Document every application, its dependencies, and its data flows. This is the input to rationalization — you can’t decide what to migrate until you know what you have.
- Application rationalization. Classify each workload: retire, retain, rehost, replatform, or refactor. Candidates for replatform or refactor (like Samsung’s Oracle-to-Aurora move) often deliver the largest cost and performance gains.
- Compliance and data residency checks. Identify regulatory requirements early. For banking-scale workloads, regulatory lead times can add 30 or more days to your cutover plan. Confirm data residency requirements for each workload before selecting AWS regions.
- Performance baselining. Measure your heaviest queries, peak traffic patterns, and latency benchmarks before migration. This is your comparison point post-cutover. Samsung’s Aurora migration validated that 85–90% of PostgreSQL queries matched performance expectations — that kind of baseline is what makes a go/no-go decision defensible.
- PoC on one high-value workflow. Target 4–12 weeks. Pick a workflow with clear business KPIs so the outcome is measurable. This is how you secure executive buy-in and validate security controls before full rollout.
- Outage window planning. For regulated workloads, coordinate downtime windows with internal stakeholders and, where required, external regulators. Build rollback runbooks before the cutover date, not during it.
- Pilot, then scale. Run a limited pilot with real traffic before full migration. Measure against your baseline. Only scale when the pilot metrics are stable.
Questions to lock scope with stakeholders: What are the business KPIs this migration must move? What’s the acceptable downtime window? Who owns regulatory coordination? What does success look like at 90 days post-migration?
For a detailed execution framework, the AWS migration best practices guide covers the full lifecycle with architecture examples.
When does it make sense to bring in an AWS migration partner?
Four signals that internal resources alone won’t be enough:
- Constrained internal capacity. If your engineering team is already running at full load, adding a migration project without external support is a schedule risk.
- Complex regulated workloads. Compliance automation, audit trail design, and regulatory coordination require specialized experience that most internal teams build only once.
- High user impact. Migrations touching customer-facing systems — payments, identity, streaming — need rollback plans that have been tested, not written.
- Multi-region data requirements. Data residency rules across jurisdictions add legal and architectural complexity that benefits from a partner who has solved it before.
What to expect from a strong partner:
- A PoC delivered on a real business workflow within an agreed timeline and budget
- Compliance automation capabilities built into the migration design
- Documented runbooks for cutover and rollback, reviewed before go-live
- Measurable cost and performance targets agreed upfront, not post-migration
Interview questions and KPIs to request: How many enterprise migrations have you completed? What AWS Partner tier are you? Can you provide references from similar verticals (fintech, eCommerce, regulated industries)? What does your rollback plan look like for a mission-critical cutover?
Red flags: A partner who won’t commit to a PoC before full engagement, or who can’t produce a written rollback plan, is not ready for a high-stakes migration.
IT-Magic operates as an AWS Advanced Tier Partner with 700+ completed migration projects, specializing in high-load eCommerce and fintech environments. For teams evaluating partners, that project volume and vertical depth are the reference points worth asking about.
What measurable outcomes do enterprise AWS migrations actually produce?
Real numbers from AWS case studies give you a benchmark to set your own targets.
| Case Study | Key Metric | Measurement Method | Business Impact |
|---|---|---|---|
| Samsung (Oracle → Aurora) | lower monthly costs | Cost comparison pre/post migration | Freed budget reallocated to product development |
| Nubank (Aurora PostgreSQL) | Up to 1,900x query improvement in specific cases; 15% average across workloads; 25% cost reduction | Query benchmarking, cost tracking | Faster payments processing, lower infrastructure spend |
| ExxonMobil (Digital Project Home) | Up to 30,000 engineering hours saved per capital project | Engineering time tracking | Direct labor cost reduction at project scale |
| BMW (EKS refactor) | AI training time cut from hours to minutes | Training job duration logs | Faster AI iteration across vehicle fleet |
Nubank’s 25% cost reduction is a useful planning benchmark for database migrations from legacy systems. That freed budget went directly into product features — which is the business case for migration, not the technical one. For teams modeling their own savings, the cost reduction guide walks through the calculation framework.
Data center reliability is another dimension worth planning for. Real-world power outage case studies show how infrastructure dependencies affect uptime during and after migration — a useful reference when designing your resilience architecture.
Key Takeaways
The most reliable path to a safe enterprise AWS migration is a PoC built around a real business workflow, followed by staged pilots with measurable baselines before full-scale rollout.
| Point | Details |
|---|---|
| Start with a business workflow | Run a 4–12 week PoC on one high-value workflow before committing to full migration scope. |
| Standardize before migrating | Reduce your application portfolio first; Mercedes-Benz’s approach shows this unlocks AI/ML capabilities post-migration. |
| Automate compliance early | Nomura built compliance agents into the migration design; regulated teams should do the same, not treat it as phase two. |
| Set measurable targets | Nubank’s 25% cost reduction is a realistic benchmark for database migrations. |
| IT-Magic as migration partner | IT-Magic is an AWS Advanced Tier Partner with 700+ completed projects, specializing in high-load eCommerce and fintech migrations. |
The mistake most enterprises make — and what actually fixes it
The single biggest migration mistake is treating the move to AWS as an IT project rather than a business transformation. That framing leads teams to optimize for technical completion — servers moved, databases replicated — while the business KPIs that justified the migration go unmeasured until someone asks six months later why costs didn’t drop.
The fix is simpler than it sounds: define the business outcome before you define the architecture. ExxonMobil saved up to 30,000 engineering hours per capital project because the migration was scoped around that specific outcome from day one. The architecture served the goal, not the other way around.
The second mistake is underestimating how much regulatory coordination costs in calendar time. For banking and payments workloads, the technical migration is often the shorter part of the project. The longer part is aligning with regulators, scheduling downtime windows, and getting sign-off on rollback procedures. Teams that discover this at week eight of a twelve-week plan don’t recover gracefully.
One practical fix: baseline your heaviest queries and peak traffic patterns before any cutover. That data is what separates a defensible go/no-go decision from a gut call. Samsung’s Aurora migration team validated 85–90% of query performance before go-live. That’s the standard worth holding yourself to.
What IT-Magic’s AWS migration service delivers for your team
Migrating to AWS without a partner is possible. Doing it without downtime, cost overruns, or compliance gaps in a high-load environment is a different challenge.
IT-Magic covers the full migration lifecycle — infrastructure audit, strategy, hands-on implementation, and post-migration optimization — with 700+ completed projects as an AWS Advanced Tier Partner. The focus is eCommerce and fintech, where a failed cutover isn’t an inconvenience; it’s lost revenue. Deliverables in the first 30–90 days include a PoC on your highest-value workflow, documented rollback runbooks, and measurable cost and performance targets agreed before work begins. For proof points, the IT-Magic case studies page shows measured outcomes across completed migrations. If you’re ready to scope your migration, start with a consultation to get a clear picture of timeline, cost, and risk before committing to full execution.
Sources and further reading
Primary case studies used in this article:
- Samsung migrates Samsung Account to Amazon Aurora
- Migrating mission-critical payments at Nubank to Amazon Aurora PostgreSQL
- BMW accelerates AI innovation on Amazon EKS
- Mercedes-Benz transforms IT operations with AWS
- ExxonMobil transforms collaboration on AWS
- Nomura uses generative AI on AWS
- AWS’s biggest customers — CloudZero
Further reading on awsmigrationservices.com:
- Migration roadmaps: secure and cost-efficient AWS moves
- AWS services guide: scalable, secure cloud migration
- IT-Magic case studies
FAQ
Which companies are the biggest AWS clients?
Netflix, Mercedes-Benz, ExxonMobil, Samsung, and Nomura Group are among the most prominent large AWS clients, each running distinct workload types from streaming infrastructure to compliance automation and global user identity systems.
How much does Netflix spend on AWS?
Netflix supports a very large number of paid subscribers and is estimated to spend hundreds of millions annually on AWS, making it one of the platform’s largest enterprise customers.
What migration strategy do large enterprises use on AWS?
Most large AWS enterprise customers use a combination of rehost, replatform, and refactor strategies, with PoC-first engagements on high-value business workflows to validate architecture and secure executive buy-in before full rollout.
How long does an enterprise AWS migration take?
Assessment typically runs several weeks, a targeted PoC takes 4–12 weeks, and a full pilot-to-scale migration spans several months depending on workload complexity, regulatory requirements, and the size of the application portfolio.
When should a business hire an AWS migration partner?
Hire a partner when internal capacity is constrained, workloads are regulated, or cutovers are customer-facing. IT-Magic, an AWS Advanced Tier Partner with 700+ completed projects, specializes in high-load eCommerce and fintech migrations where downtime directly affects revenue.