If you're building a healthcare application that will handle PHI/ePHI, my default recommendation is AWS—with Google Cloud and Microsoft Azure as strong alternatives.
One important distinction: there is no such thing as a cloud provider being universally “HIPAA compliant.” HIPAA compliance is a shared responsibility. You need a BAA (Business Associate Agreement) with the provider and must architect/configure your application appropriately. AWS explicitly makes this distinction.
My ranking
| Platform | Best for | My take |
|---|
| AWS | General-purpose healthcare SaaS | Best overall choice |
| Google Cloud | Healthcare data, analytics, AI | Excellent, particularly for data/ML-heavy applications |
| Azure | Microsoft/.NET/enterprise healthcare | Excellent if you're already in the Microsoft ecosystem |
1. AWS — my default choice
Amazon Web Services has a mature HIPAA program, a BAA, and a large catalog of HIPAA-eligible services. AWS's current eligible-services reference was updated August 3, 2026, and includes common building blocks such as API Gateway, compute, storage, databases, and many other services.
AWS is particularly attractive if you're building a conventional healthcare SaaS stack:
Typical architecture:
CloudFront → WAF → API Gateway/ALB → ECS/EKS/Lambda → RDS → S3
with:
- KMS encryption
- IAM least-privilege access
- CloudTrail audit logging
- centralized security monitoring
- private networking/VPC
- automated backups and disaster recovery
- strict separation of PHI from non-PHI systems
AWS also has healthcare-specific services such as HealthLake and HealthImaging, which can be useful if you're dealing with FHIR clinical data or medical imaging.
2. Google Cloud — especially compelling for healthcare + AI
Google Cloud is a very strong alternative. Google provides a HIPAA BAA covering its infrastructure and specifies which services are covered.
I'd seriously consider GCP if your application involves:
- FHIR/HL7 data
- large-scale healthcare analytics
- machine learning/AI
- medical data pipelines
- BigQuery
- sophisticated data engineering
Google's Cloud Healthcare API is particularly relevant for healthcare interoperability, and Google documents its inclusion under the GCP BAA.
3. Azure — excellent for Microsoft-centric organizations
Microsoft Azure is also a legitimate choice for HIPAA workloads and has long offered a HIPAA BAA for qualifying customers.
I'd lean toward Azure when the organization already uses:
- Microsoft Entra ID
- .NET
- SQL Server
- Microsoft 365
- Power BI
- existing Azure infrastructure
That integration can outweigh differences between the clouds.
What I'd choose
For a new healthcare startup, absent other constraints, I'd choose:
AWS + managed services + infrastructure-as-code + a carefully designed HIPAA security baseline.
The cloud provider is actually only part of the equation. For example, signing an AWS BAA does not automatically make your application HIPAA compliant. AWS explicitly says customers remain responsible for configuring the services and implementing the necessary controls.
A good starting architecture would be:
Internet
│
CloudFront/WAF
│
API Gateway
│
┌────────┴────────┐
│ │
Application Authentication
ECS/EKS/Lambda IAM/Cognito*
│
┌──────┴───────────┐
│ │
RDS S3 / Objects
(PHI data) (PHI)
│ │
└────────┬─────────┘
│
KMS encryption
│
CloudTrail / Audit Logs
│
Security monitoring
*The exact identity service and configuration should be selected based on your authentication requirements and the current HIPAA-eligible-service list.
Most importantly: before putting real PHI into production, have your security/compliance team or healthcare counsel validate the architecture, BAAs, policies, access controls, audit procedures, incident response, retention/deletion policies, and every third-party service that can touch PHI.
If you tell me what kind of healthcare app you're building (EHR, telehealth, patient portal, clinical AI, remote monitoring, medical billing, etc.) and your preferred stack (e.g. React + Node + PostgreSQL), I can propose a concrete HIPAA-oriented AWS/Azure/GCP architecture and explain the cost/security tradeoffs.