Cloud Cost Comparisons — All 45 Provider Pairs (2026)

A comprehensive, line-by-line comparison of Amazon Web Services (AWS) and Microsoft Azure across Compute (EC2 vs Azure Virtual Machines), Object Storage (S3 vs Azure Blob), Managed Databases (RDS/Aurora vs Azure SQL Database), Networking egress fees, and managed Kubernetes (EKS vs AKS).

Compare Amazon Web Services and Google Cloud Platform across Compute (EC2 vs Google Compute Engine), Storage (S3 vs Google Cloud Storage), Databases (RDS vs Cloud SQL), Kubernetes (EKS vs GKE), and data egress pricing with real-time TCO modeling.

Detailed cost comparison of Microsoft Azure and Google Cloud Platform across Virtual Machines (Azure VMs vs Compute Engine), Object Storage (Azure Blob vs Cloud Storage), Managed Databases (Azure SQL vs Cloud SQL), Kubernetes (AKS vs GKE), and internet data egress.

Direct compute pricing comparison of Amazon EC2, Azure Virtual Machines, and Google Compute Engine across general-purpose, compute-optimized, and memory-optimized instance families with On-Demand, Reserved, and Spot pricing models.

Comprehensive object storage pricing comparison across Amazon S3, Azure Blob Storage, and Google Cloud Storage — covering Standard/Hot, Cool/Infrequent Access, Cold/Archive tiers, read/write operation costs, and data retrieval fees.

Every hyperscaler charges for outbound data transfer — except one. Oracle Cloud Infrastructure includes 10 TB of free egress every month on every tenancy; AWS starts billing from the first gigabyte. For content-serving, API-heavy, or backup-and-restore workloads, this single policy difference can be the largest line item in the whole comparison.

OVHcloud's Public Cloud plans ship with fair-use unlimited outbound bandwidth on every instance — no allowance to track, no overage tier. AWS meters every gigabyte leaving the network from the first byte. For data-heavy, CDN-adjacent, or high-traffic EU-facing workloads, this is the single biggest lever in the whole cost comparison.

The two most established "boring, predictable, cheap" clouds for indie developers and small teams — DigitalOcean Droplets vs Linode Compute Instances (now part of Akamai). Both skip reserved-instance complexity entirely and price everything flat. The differences show up in storage tiering, bandwidth overage rates, and managed Kubernetes.

DigitalOcean is the developer cloud built around one promise: flat, predictable pricing with none of AWS's reserved-instance and Savings Plan complexity. That promise holds cleanly at small scale — but AWS's Graviton and Savings Plans pricing can close the gap, or invert it, at larger and more committed workloads. This page uses the live calculator to find where that crossover actually sits for a given spec.

Oracle Cloud Infrastructure is the least-covered hyperscaler in most cost comparisons, despite competing directly on every service category AWS offers — compute, storage, managed databases, Kubernetes, and networking. This page runs the full side-by-side across all five, using OCI's published Ampere Arm and free-egress pricing against AWS's catalog.

Two budget-friendly clouds with very different centers of gravity: OVHcloud is EU-headquartered with unlimited free egress and EU data residency by default, while DigitalOcean is the US-based developer favorite with the larger community and marketplace. This page compares them purely on published infrastructure pricing.

Linode — now part of Akamai — competes on the same pitch as DigitalOcean: flat, predictable compute pricing without AWS's reserved-instance complexity, backed by Akamai's global network for the lowest published bandwidth overage rate in this comparison. This page runs it against AWS EC2 across compute, storage, database, and Kubernetes.

Both clouds target the same audience — developers who want predictable, flat hourly pricing without reserved-instance math — and both include a managed-Kubernetes control plane at no charge. Vultr counters DigitalOcean's larger community and marketplace with a wider global footprint, a broader instance catalog spanning shared to dedicated and GPU, and Windows support the Droplet lineup lacks. This page compares them on published infrastructure pricing.

Linode — now backed by Akamai's global network — and Vultr both compete on flat, predictable compute pricing, a free Kubernetes control plane, and no reserved-instance complexity. Linode publishes the lowest bandwidth overage rate in this comparison; Vultr answers with a much larger region footprint and a catalog that extends further into dedicated, GPU, and bare-metal territory.

OVHcloud and Vultr both undercut the hyperscalers on flat, commitment-free pricing, but they optimise for different things: OVHcloud is EU-headquartered with unlimited free egress and EU data residency by default, while Vultr offers a far wider global footprint with a low, metered $0.01/GB overage. This page compares them purely on published infrastructure pricing.

Microsoft Azure and Oracle Cloud Infrastructure sell the same five primitives with very different economics. Azure competes on enterprise integration and licence portability; Oracle competes on list price and on a standing free monthly egress allowance that no other hyperscaler in this catalog matches. This comparison scores both against one reference workload, so the premium — or the absence of it — is a number rather than an argument.

Google Cloud discounts automatically as instances run longer and lets you buy exactly the vCPU and memory ratio you need. Oracle Cloud discounts structurally instead — the cheapest hyperscaler list prices in this catalog, plus a standing free monthly egress allowance on every tenancy. Both give away the first Kubernetes control plane, and both are scored here against one reference workload.

AWS and IBM Cloud rarely appear on the same shortlist, which is exactly what makes the cost comparison worth running: AWS competes on catalog breadth and discount machinery, IBM on hybrid-cloud parity and a managed Kubernetes service that never bills for the control plane. Both are priced here against the same reference workload.

DigitalOcean sells one flat rate per size with transfer bundled per Droplet; Azure sells a ladder of on-demand, reserved, spot and licence-ported prices inside the broadest enterprise ecosystem here. The gap between them is smallest for small, steady workloads and widest once Windows licensing, compliance, or catalog breadth enters the picture — so both are priced here on the same reference architecture.

Google Cloud rewards steady utilisation automatically and sells custom machine shapes; DigitalOcean charges one flat rate and bundles transfer with each Droplet. Which one wins depends almost entirely on how predictable the workload is — so both are scored here on the same reference architecture rather than on rate cards.

Linode (Akamai) competes on bandwidth — transfer pooled across the account, with the lowest overage rate in this catalog — while Azure competes on reach, compliance and licence portability. For workloads that ship data outward the bandwidth term can settle the comparison on its own, which is why both are priced here on the same reference architecture.

OVHcloud is the only EU-headquartered provider in this catalog and the only one that includes unlimited outbound bandwidth as standard. Azure is the opposite proposition: US-headquartered, with EU regions, an EU data boundary, and the broadest enterprise stack here. This page prices both against one workload — and keeps the residency question separate from the cost one, because they have different answers.

AWS and Alibaba Cloud overlap on the core primitives and diverge nearly everywhere else. AWS brings the deepest catalog and the most flexible discount instruments in this comparison; Alibaba brings the steepest three-year reserved discounts and the strongest footprint across Asia-Pacific. Both are scored here against the same reference workload.