August 6, 2026
Cloud Pricing Trends: August 2026
Cloud prices aren't static — and most simulators treat them as if they are. We maintain a verified, provider-by-provider pricing history so our simulations are always grounded in what you'd actually be charged today.
Why we track pricing history
Simulation accuracy starts with correct inputs. If the hourly rate for an RDS instance in our model drifts from the published AWS price, every cost estimate we produce is wrong — and it compounds across architectures. A 5% error on a single resource becomes a 20% error on a multi-tier stack that uses eight of them.
So we verify rates against official provider sources — the AWS Pricing API, the GCP Cloud Billing Catalog API, the Azure Retail Prices API, OCI's Pricing API, and DigitalOcean's Sizes API — on a monthly cadence. When a rate changes, we add a new snapshot to our pricing history rather than silently overwriting the old one. That record lets us show you trends, not just current prices.
We publish this history openly at /pricing-history. If you're evaluating a cloud migration decision or modeling a multi-year infrastructure budget, that page shows exactly how prices have moved for each shape we track.
What we track
Our history covers five providers: AWS (us-east-1), GCP (us-central1), Azure (East US), OCI (us-ashburn-1), and DigitalOcean (nyc3). Across those providers we track 33 canonical resources across four categories: compute, database, storage, and network.
We use the same shapes each snapshot — t2.micro, m5.large, m5.xlarge, e2-medium, n2-standard-4, Standard_B2s, Standard_D4s_v3, VM.Standard.E4.Flex, and so on — so comparisons are apples-to-apples across time. If a provider changes its pricing model (switching from OCPU-based to ECPU-based billing, for example), we document the methodology change alongside the rate so neither number is misread as a pure price movement.
The headline: stability, with a few real exceptions
Our pricing history now covers four snapshots: November 2025, May 2026, June 2026, and August 2026. The most striking finding is how little has actually moved. Across 33 tracked resources and five providers, the vast majority of rates are identical today to what they were nine months ago. The confirmed provider movement in these snapshots is the GCP database price reduction. The OCI VM.Standard3.Flex records differ in rate and configuration, so they do not establish a like-for-like price change. DigitalOcean's tracked Droplet prices were essentially flat (c-4 rose about 1.2%).
Provider snapshots
AWS — compute stable; one database snapshot differs
AWS compute rates have been stable throughout our tracking period: t2.micro at $0.0116/hr, m5.large at $0.096/hr, and m5.xlarge at $0.192/hr. The db.t3.small history lists $0.034/hr in November 2025, $0.036/hr in May 2026, then $0.034/hr from June through September. That snapshot variation does not, by itself, establish an AWS price movement. The other tracked AWS rates — including db.r5.large at $0.24/hr and the Application Load Balancer at $0.0225/hr — are unchanged.
GCP — stable compute; real database cost reduction
GCP compute rates have been flat throughout our tracking period: e2-medium at $0.0553/hr (2 vCPU × $0.021811 + 4 GB × $0.002923, verified via the Cloud Billing Catalog API), n2-standard-4 at $0.194/hr, and Cloud SQL db-f1-micro at $0.0105/hr. The n2-standard-4 and load balancer ($0.025/hr per forwarding rule) are also unchanged.
The one real movement: Cloud SQL db-n1-standard-2 dropped from $0.184/hr in November 2025 to $0.1351/hr starting in June 2026. This is a genuine price reduction driven by GCP's introduction of a bundled “2 vCPU + 7.5 GB” SKU, confirmed via the Cloud Billing Catalog API (service 9662-B51E-5089). Teams running Cloud SQL at that tier are paying meaningfully less than they were a year ago. That rate holds in August.
Azure — stable throughout
Azure prices have been consistent across all snapshots. Standard_B2s holds at $0.0416/hr, Standard_D4s_v3 at $0.192/hr (verified August 2026 against the Azure Retail Prices API), SQL Basic at $0.0068/hr (~$4.90/mo), and SQL Standard S2 at $0.1008/hr (~$73.58/mo). Blob Storage ($0.018/hr) and the Load Balancer ($0.025/hr) are also unchanged. No rate movements detected in August.
OCI — VM snapshots use different configurations; billing model note for Autonomous DB
The VM.Standard3.Flex snapshots list $0.045/hr for 4 OCPUs in November 2025 and $0.049/hr for 1 OCPU + 6 GB in May 2026. Because these records describe different configurations, they are not a like-for-like comparison and do not establish a provider price increase. The E4.Flex reference config — 2 OCPU + 16 GB — remains $0.074/hr. Other tracked AWS, Azure, and GCP compute rates are unchanged. Object Storage, Block Volume, File Storage, and the Load Balancer are also unchanged.
One figure warrants context: Autonomous Database at $0.789/hr. This reflects OCI's ECPU-based billing — our reference config is 2 ECPUs at $0.336/ECPU-hr plus storage. OCI migrated Autonomous DB from OCPU to ECPU billing in 2024, so any comparison against pre-migration figures needs to account for the different billing unit, not just the rate. Our snapshots all use the ECPU config consistently, so the trend line is apples-to-apples.
DigitalOcean — an apparent compute spike was a billing-basis error
A later source audit found that the reported June increases came from multiplying DigitalOcean's API price_hourly by 730, rather than using its published monthly price. The hourly field uses a different billing-hour basis, so those values did not establish price changes. The published monthly prices are $6/mo for s-1vcpu-1gb, $24/mo for s-2vcpu-4gb, and $84/mo for c-4; checks now compare each monthly price divided by 730. The managed database Basic reference plans are listed at $15.15/mo and $60.90/mo. Missing or unmatched catalog plans are not assigned substitute prices.
What this means for cloud budgets
Tracked compute prices are mostly stable. If you built a cost model against these providers nine months ago using accurate list prices, it is likely still useful, but the confirmed GCP database movement and historical snapshot differences below deserve review.
If you run Cloud SQL db-n1-standard-2 on GCP, you should be seeing lower bills since June. The $0.184 → $0.1351/hr reduction is a 27% drop for that specific tier. If your cost model hasn't been updated, it's overstating your GCP database spend.
For AWS db.t3.small, the history lists $0.034/hr in November 2025, $0.036/hr in May 2026, and $0.034/hr from June through September. The May snapshot differs, but this record alone does not establish an AWS price movement.
If you run DigitalOcean Droplets, DigitalOcean's tracked Droplet prices were essentially flat (c-4 rose about 1.2%). The tracked s-1vcpu-1gb and s-2vcpu-4gb prices did not change from November 2025 through September 2026.
For OCI VM.Standard3.Flex, the history lists $0.045/hr for 4 OCPUs in November 2025 and $0.049/hr for 1 OCPU + 6 GB in May 2026. Those configurations are not comparable, so the snapshots do not establish a provider price increase.
For AWS, Azure, and GCP compute, tracked list prices have been flat. That stability is useful for long-horizon planning: you can model a 12-month infrastructure budget with reasonable confidence that today's rates will still be the rates you pay next quarter.
The simulation lets you model the cost impact of any of these movements against your specific architecture — not a generic benchmark. If your stack runs eight m5.large instances alongside Cloud SQL, you can update to current rates and see exactly what changed and what didn't.
We update our pricing history each month and publish the full snapshot record at /pricing-history. To model what the current rates mean for your architecture, start a simulation at /simulate.
View the full pricing history →Model your architecture →
Next: Cloud Pricing Trends, September 2026