Workload-first · five stacks · open methodology

One workload.
Five honest bills.

Model requests, active CPU, wall time, storage, delivery, growth, and launch spikes once. The engine translates them into each provider’s actual billing units.

Cloudflare Workers + R2
AWS Lambda + S3
AWS EC2 fixed stack
Google Cloud Run + Cloud Storage
Hetzner self-managed VPS

Integermicrodollar math

720hourly traffic buckets

Everyrate source-linked

Zerodata sent to a server

theserverless.dev

Cloud Cost ModelerGenerated 2026-08-09

Workload estimate · schema v1

Growing SaaS API

Lowest comparable billHetzner$23.59 / month

Monthly comparison

USD infrastructure list prices. Configurations that cannot serve the entered workload are excluded from the lowest-bill result.

Provider / modeled stack	Status	Price book	Monthly total
CloudflareWorkers Paid + R2	Comparable	2026-07-24.1	$24.00
AWSLambda x86 + HTTP API + S3	Comparable	2026-07-24.1-us-east-1	$164.45
AWSFixed c6a.large fleet + ALB + S3	Comparable · cost only	2026-07-24.1-us-east-1	$259.41
Google CloudCloud Run request-based + Cloud Storage	Comparable	2026-07-24.1-us-central1	$274.56
HetznerFixed CPX22 self-managed VPS	Comparable · cost onlyLowest comparable	2026-07-24.1-eu	$23.59

Inputs and assumptions

The report records the workload and capacity assumptions that produced these totals.

Monthly requests: 50M
Active CPU / request: 7 ms
Wall time / request: 35 ms
Required memory: 128 MiB
Inbound payload: 2 KB
Public response: 25 KB
Stored data: 50 GB-month
Object reads: 0
Object writes: 0
Traffic shape: business hours
Monthly growth: 8%
Launch spike: 5× for 6 h
Average traffic: 19.29 rps
Peak traffic: 174.48 rps
Peak / average: 9.05×
Free allowances: Available (greenfield)
AWS Lambda ingress: API Gateway HTTP API
AWS EC2 reference: 2 × c6a.large · 65% target
Google Cloud Run: 1 vCPU · concurrency 1
Hetzner reference: 1 × CPX22 · 65% target

Growth sensitivity

Request growth compounds monthly; behavior and infrastructure settings stay fixed. This is not a forecast.

Provider	Month 1	Month 12	Month 36	Qualification
Cloudflare	$24.00	$53.30	$327.28	Comparable through month 36
AWS	$164.45	$395.99	$2,504	Comparable through month 36
AWS	$259.41	$423.60	$1,917	Partial cost from month 12
Google Cloud	$274.56	$633.60	$3,782	Comparable through month 36
Hetzner	$23.59	$23.59	$23.59	Partial cost from month 3

Scope and verification

List-price estimate, not a quote. Lowest list price is not a recommendation. Capability, capacity, availability, support, operations, taxes, negotiated discounts, observability, database services, NAT, and other architecture choices can materially change a real bill.

Price entries checked 2026-07-24. Detailed official sources follow on each provider page. Review the model at theserverless.dev/methodology and the normalized prices at theserverless.dev/pricing-data.

04 · Provider detail

Cloudflare

Global edge runtime; active CPU billing

Comparable$24.00/month · 2026-07-24.1

Average19.29 rps

Peak174.48 rps

Modeled capacityElastic reference

Top cost driverWorker requests

Itemized monthly bill

Item / rule / source	Quantity / applied rate	Billing formula	Cost
Workers Paid account minimumM-CF-001 · checked 2026-07-24 · developers.cloudflare.com	1 account-month$5.00 / unit	1 Workers Paid account minimumOne account-level minimum; not a per-application platform fee.	$5.00
Worker requestsM-CF-001 · checked 2026-07-24 · developers.cloudflare.com	50M requests$0.3000 / 1M units	40,000,000 requests after 10,000,000 account allowance	$12.00
Worker CPU timeM-CF-001 · checked 2026-07-24 · developers.cloudflare.com	350M CPU-ms$0.0200 / 1M units	320,000,000 active CPU-ms after 30,000,000 account allowanceWorkers bills active CPU, not wall-clock time spent waiting on I/O.	$6.40
R2 Standard storageM-CF-002 · checked 2026-07-24 · developers.cloudflare.com	50 GB-month$0.0150 / unit	40 billed after 10 recurring allowance	$0.60
R2 Class A operationsM-CF-002 · checked 2026-07-24 · developers.cloudflare.com	0 operationsno billed usage	0 billed write operations	$0.00
R2 Class B operationsM-CF-002 · checked 2026-07-24 · developers.cloudflare.com	0 operationsno billed usage	0 billed read operations	$0.00
Internet egressM-CF-003 · checked 2026-07-24 · developers.cloudflare.com	1.3K GB$0 / unit	1250.000 GB × $0 internet egressExplicit zero-dollar line: R2 internet egress is not charged.	$0.00
Monthly total			$24.00

Model assumptions

Responses are served through Workers and object data is held in R2 Standard.
Active CPU time is supplied separately from wall time.
R2 operation counts are supplied explicitly; no database, queue or observability charges are inferred.

Model notes and limits

shared allowanceGreenfield view assumes recurring account-wide free allowances are unused. Promotional credits are excluded.

05 · Provider detail

AWS

us-east-1 list price; wall-time compute billing

Comparable$164.45/month · 2026-07-24.1-us-east-1

Average19.29 rps

Peak174.48 rps

Modeled capacityElastic reference

Top cost driverInternet egress (first 10 TB tier)

Itemized monthly bill

Item / rule / source	Quantity / applied rate	Billing formula	Cost
Lambda requestsM-AWS-LAMBDA-001 · checked 2026-07-24 · aws.amazon.com	50M requests$0.2000 / 1M units	49,000,000 invocations after 1,000,000 shared allowance	$9.80
Lambda x86 computeM-AWS-LAMBDA-001 · checked 2026-07-24 · aws.amazon.com	218.8K GB-secondsno applied charge	50,000,000 × 35 ms (1 ms rounded) × 0.125 GB; 400,000 GB-s allowanceLambda bills allocated memory for wall-clock execution, including I/O wait.	$0.00
API Gateway HTTP APIM-AWS-APIGW-001 · checked 2026-07-24 · aws.amazon.com	50M metered requests$1.00 / 1M units	50,000,000 incoming requests × 1 request unit(s) at 512 KB eachThe 512 KB multiplier applies to inbound request payload, not response egress.	$50.00
API Gateway HTTP API (above 300M)M-AWS-APIGW-001 · checked 2026-07-24 · aws.amazon.com	50M metered requestsno applied charge	0 metered requests above the first 300M	$0.00
S3 Standard storageM-AWS-S3-001 · checked 2026-07-24 · aws.amazon.com	50 GB-month$0.0230 / unit	50 billed after 0 recurring allowance	$1.15
S3 PUT/COPY/POST/LISTM-AWS-S3-001 · checked 2026-07-24 · aws.amazon.com	0 requestsno billed usage	0 billed write operations	$0.00
S3 GET/SELECTM-AWS-S3-001 · checked 2026-07-24 · aws.amazon.com	0 requestsno billed usage	0 billed read operations	$0.00
Internet egress (first 10 TB tier)M-AWS-NET-001 · checked 2026-07-24 · aws.amazon.com	1.3K GB$0.0900 / unit	1250.000 GB total − 100 GB shared allowance; 1150.000 GB in this tier	$103.50
Internet egress (next 40 TB tier)M-AWS-NET-001 · checked 2026-07-24 · aws.amazon.com	1.3K GBno applied charge	0.000 GB in this tier	$0.00
Internet egress (next 100 TB tier)M-AWS-NET-001 · checked 2026-07-24 · aws.amazon.com	1.3K GBno applied charge	0.000 GB in this tier	$0.00
Internet egress (next 350 TB tier)M-AWS-NET-001 · checked 2026-07-24 · aws.amazon.com	1.3K GBno applied charge	0.000 GB in this tier	$0.00
Monthly total			$164.45

Model assumptions

Lambda x86 in us-east-1 with 1 ms duration rounding.
The entered wall time includes initialization and I/O wait.
S3 Standard is modeled without CloudFront, transfer acceleration, replication or lifecycle retrieval fees.
CloudWatch, X-Ray, build artifacts, database, NAT and support charges are excluded.

Model notes and limits

shared allowanceGreenfield view assumes recurring account-wide free allowances are unused. Promotional credits are excluded.

06 · Provider detail

AWS

Capacity-qualified fixed-instance reference

Comparable · cost only$259.41/month · 2026-07-24.1-us-east-1

Average19.29 rps

Peak174.48 rps

Modeled capacity390 rps

Top cost driverEC2 c6a.large Linux On-Demand

Itemized monthly bill

Item / rule / source	Quantity / applied rate	Billing formula	Cost
EC2 c6a.large Linux On-DemandM-AWS-EC2-001 · checked 2026-07-24 · aws.amazon.com	1.4K instance-hour$0.0765 / unit	2 fixed c6a.large node(s) × 720 hours2 vCPU, 4 GiB, shared-nothing application node reference.	$110.16
EBS gp3 root volumeM-AWS-EC2-001 · checked 2026-07-24 · aws.amazon.com	40 GB-month$0.0800 / unit	2 node(s) × 20 GB gp3 root volumeIncludes gp3 baseline performance; snapshots are excluded.	$3.20
Application Load BalancerM-AWS-ALB-001 · checked 2026-07-24 · aws.amazon.com	720 load-balancer-hour$0.0225 / unit	720 load-balancer hours	$16.20
Application Load Balancer capacityM-AWS-ALB-001 · checked 2026-07-24 · aws.amazon.com	1.3K LCU-hour$8.00 / 1K units	1350.000 LCU-hours; hourly maximum of new connections, active connections, bytes and rule evaluations	$10.80
Public IPv4 addressesM-AWS-EC2-001 · checked 2026-07-24 · aws.amazon.com	2.9K address-hour$5.00 / 1K units	(2 internet-facing ALB address(es) + 2 public node address(es)) × 720 hours	$14.40
S3 Standard storageM-AWS-S3-001 · checked 2026-07-24 · aws.amazon.com	50 GB-month$0.0230 / unit	50 billed after 0 recurring allowance	$1.15
S3 PUT/COPY/POST/LISTM-AWS-S3-001 · checked 2026-07-24 · aws.amazon.com	0 requestsno billed usage	0 billed write operations	$0.00
S3 GET/SELECTM-AWS-S3-001 · checked 2026-07-24 · aws.amazon.com	0 requestsno billed usage	0 billed read operations	$0.00
Internet egress (first 10 TB tier)M-AWS-NET-001 · checked 2026-07-24 · aws.amazon.com	1.3K GB$0.0900 / unit	1250.000 GB total − 100 GB shared allowance; 1150.000 GB in this tier	$103.50
Internet egress (next 40 TB tier)M-AWS-NET-001 · checked 2026-07-24 · aws.amazon.com	1.3K GBno applied charge	0.000 GB in this tier	$0.00
Internet egress (next 100 TB tier)M-AWS-NET-001 · checked 2026-07-24 · aws.amazon.com	1.3K GBno applied charge	0.000 GB in this tier	$0.00
Internet egress (next 350 TB tier)M-AWS-NET-001 · checked 2026-07-24 · aws.amazon.com	1.3K GBno applied charge	0.000 GB in this tier	$0.00
Operations, patching and observabilityM-AWS-EC2-002 · checked 2026-07-24 · aws.amazon.com	0 engineer-hoursno billed usage	Not inferable from a workload aggregateA $0 line is not a claim that operating instances is free.	$0.00
Monthly total			$259.41

Model assumptions

Linux On-Demand c6a.large in us-east-1, billed for a fixed 720-hour month.
Application nodes are stateless; persistent objects use S3 Standard.
ALB LCU-hours are integrated from the 720-bucket traffic curve in HA mode.
Cross-AZ transfer, NAT gateways, database, snapshots, logs, support and human operations are excluded.

Model notes and limits

shared allowanceGreenfield view assumes recurring account-wide free allowances are unused. Promotional credits are excluded.
metering assumption300 RPS/node at 65% target is your benchmark assumption, not an AWS performance guarantee.

07 · Provider detail

Google Cloud

us-central1; concurrency 1, min instances 0

Comparable$274.56/month · 2026-07-24.1-us-central1

Average19.29 rps

Peak174.48 rps

Modeled capacityElastic reference

Top cost driverPremium Tier egress (up to 1 TiB)

Itemized monthly bill

Item / rule / source	Quantity / applied rate	Billing formula	Cost
Cloud Run request CPUM-GCP-RUN-001 · checked 2026-07-24 · cloud.google.com	5M vCPU-seconds$24.00 / 1M units	50,000,000 × 0.1 s (100 ms rounded) × 1 vCPU; 180,000 vCPU-s allowanceRequest-based billing allocates CPU for the billable request lifetime.	$115.68
Cloud Run request memoryM-GCP-RUN-001 · checked 2026-07-24 · cloud.google.com	625K GiB-seconds$2.50 / 1M units	50,000,000 × 0.1 s × 0.125 GiB; 360,000 GiB-s allowance	$0.66
Cloud Run requestsM-GCP-RUN-001 · checked 2026-07-24 · cloud.google.com	50M requests$0.4000 / 1M units	48,000,000 requests after 2,000,000 allowance	$19.20
Cloud Storage StandardM-GCP-STORAGE-001 · checked 2026-07-24 · cloud.google.com	46.6 GiB-month$0.0200 / unit	41.566 billed after 5 recurring allowance	$0.83
Cloud Storage Class A operationsM-GCP-STORAGE-001 · checked 2026-07-24 · cloud.google.com	0 operationsno billed usage	0 billed write operations	$0.00
Cloud Storage Class B operationsM-GCP-STORAGE-001 · checked 2026-07-24 · cloud.google.com	0 operationsno billed usage	0 billed read operations	$0.00
Premium Tier egress (up to 1 TiB)M-GCP-NET-001 · checked 2026-07-24 · cloud.google.com	1.2K GiB$0.1200 / unit	1164.153 GiB total − 1 GiB allowance; 1024.000 GiB in this tier	$122.88
Premium Tier egress (1–10 TiB)M-GCP-NET-001 · checked 2026-07-24 · cloud.google.com	1.2K GiB$0.1100 / unit	139.153 GiB from 1–10 TiB	$15.31
Premium Tier egress (above 10 TiB)M-GCP-NET-001 · checked 2026-07-24 · cloud.google.com	1.2K GiBno applied charge	0.000 GiB above 10 TiB	$0.00
Monthly total			$274.56

Model assumptions

Cloud Run request-based billing in us-central1, one request per instance at a time.
Wall time includes weighted startup and shutdown time; duration rounds to 100 ms.
Cloud Storage quantities are converted from decimal GB to binary GiB.
Cloud Run serves public response bytes; CDN, build, artifact registry, logs, database and VPC products are excluded.
Cloud Run is a regional, zonally redundant managed service; no extra HA replica count is inferred.

Model notes and limits

shared allowanceGreenfield view assumes recurring account-wide free allowances are unused. Promotional credits are excluded.

08 · Provider detail

Hetzner

EU list-price reference; excludes VAT and labor

Comparable · cost only$23.59/month · 2026-07-24.1-eu

Average19.29 rps

Peak174.48 rps

Modeled capacity195 rps

Top cost driverHetzner CPX22

Itemized monthly bill

Item / rule / source	Quantity / applied rate	Billing formula	Cost
Hetzner CPX22M-HETZNER-001 · checked 2026-07-24 · docs.hetzner.com	1 server-month$22.99 / unit	1 fixed CPX22 server(s)Shared-resource reference; local taxes are excluded.	$22.99
Primary IPv4M-HETZNER-001 · checked 2026-07-24 · docs.hetzner.com	1 address-month$0.6000 / unit	1 primary IPv4 address(es)	$0.60
Server backupsM-HETZNER-001 · checked 2026-07-24 · docs.hetzner.com	0 server-monthno billed usage	Backups disabled	$0.00
Included local NVMe storageM-HETZNER-002 · checked 2026-07-24 · docs.hetzner.com	50 GB-monthno applied charge	50.000 GB requested; 80 GB included across nodesLocal storage is not object storage and is not replicated by this model.	$0.00
Traffic above included allowanceM-HETZNER-003 · checked 2026-07-24 · docs.hetzner.com	1.3 TBno applied charge	1250.000 GB total − 20,000 GB included; 0.000 TB billable	$0.00
Self-managed operationsM-HETZNER-004 · checked 2026-07-24 · docs.hetzner.com	0 engineer-hoursno billed usage	Self-managed OS, runtime, TLS, scaling, backups and incident responseList-price estimate only. Human operations and downtime risk are not monetized.	$0.00
Monthly total			$23.59

Model assumptions

CPX22 in Germany or Finland, using the published post-15-June-2026 USD list price.
The workload fits the included local disk and accepts non-replicated local persistence.
20 TB outbound traffic per node is treated as a pooled comparison allowance.
Database, load balancing, snapshots beyond the backup toggle and human operations are excluded.

Model notes and limits

metering assumption300 RPS/node at 65% target is your benchmark assumption. CPX22 uses shared resources.
metering assumptionThis is a self-managed reference: application deployment, TLS, monitoring, security updates and on-call work are not priced.

Browser-local modelPrices checked 24 Jul 2026

01 · Define workload

Growing SaaS API

Model name

Request compute

Separate active CPU from end-to-end duration. Providers meter these differently.

Monthly requestsProvider-neutral application operations.requests / monthActive CPU timeTime actually executing application code.ms / requestWall-clock durationIncludes I/O wait and upstream latency.ms / requestRequired working memoryMinimum memory needed per execution; hard provider limits can disqualify a stack.MiBInbound payloadUsed for request metering where applicable.KB / requestPublic responseDrives modeled Internet egress.KB / response

Object storage

One storage workload mapped to R2, S3, Cloud Storage, or explicit gaps.

Stored dataAverage storage for the billing month.GB-monthObject readsGET-like operations during the month.operations / monthObject writesPUT/list/state-changing operations.operations / month

Allowance view

Greenfield · allowances availableMarginal · allowances consumed

AWS request ingress

API Gateway HTTP APILambda Function URL

03 · Selected provider bill

Cloudflare

Global edge runtime; active CPU billing

Comparable monthly total$24.002026-07-24.1

Item / rule	Quantity	Formula	Cost
Workers Paid account minimum M-CF-001 · checked 2026-07-24	1 account-month	1 Workers Paid account minimumOne account-level minimum; not a per-application platform fee.	$5.00
Worker requests M-CF-001 · checked 2026-07-24	50M requests	40,000,000 requests after 10,000,000 account allowance	$12.00
Worker CPU time M-CF-001 · checked 2026-07-24	350M CPU-ms	320,000,000 active CPU-ms after 30,000,000 account allowanceWorkers bills active CPU, not wall-clock time spent waiting on I/O.	$6.40
R2 Standard storage M-CF-002 · checked 2026-07-24	50 GB-month	40 billed after 10 recurring allowance	$0.60
Monthly total			$24.00

Capacity

Average: 19.29 rps
Peak: 174.48 rps
Comparison status: Comparable

Assumptions

Responses are served through Workers and object data is held in R2 Standard.
Active CPU time is supplied separately from wall time.
R2 operation counts are supplied explicitly; no database, queue or observability charges are inferred.

Model notes and limits

shared allowanceGreenfield view assumes recurring account-wide free allowances are unused. Promotional credits are excluded.

04 · Look ahead

Growth changes the answer.

The projection compounds request volume while holding per-request behavior and infrastructure assumptions constant. It is a sensitivity test, not a forecast.

36-month run rate$327.28 / month

At 8% monthly growth, volume reaches 739.3M requests.

Configured stack remains comparable through month 36.

M1M6M12M18M24M30M36

Request-volume sensitivityCloudflare

0.25× traffic12.5M req$7.50

0.5× traffic25M req$13.00

1× traffic50M req$24.00

2× traffic100M req$46.00

4× traffic200M req$90.00

List-price comparison, not a quote. Taxes, negotiated discounts, free-tier eligibility, support, observability, database choices, NAT, and human operations can materially change a real bill.

Challenge the methodology Inspect pricing data

The important caveat

Lowest price is not the same as best architecture.

A fixed server can look spectacularly cheap while being too small for the modeled peak. A managed stack can cost more while removing operations work. The model keeps those facts separate: price, capacity eligibility, capability gaps, and assumptions.

Read methodology Inspect pricing data

One workload.Five honest bills.

Growing SaaS API

Monthly comparison

Inputs and assumptions

Growth sensitivity

Cloudflare

Itemized monthly bill

Model assumptions

Model notes and limits

AWS

Itemized monthly bill

Model assumptions

Model notes and limits

AWS

Itemized monthly bill

Model assumptions

Model notes and limits

Google Cloud

Itemized monthly bill

Model assumptions

Model notes and limits

Hetzner

Itemized monthly bill

Model assumptions

Model notes and limits

Growing SaaS API

Request compute

Object storage

Cloudflare

Line items

Capacity

Assumptions

Model notes and limits

Growth changes the answer.

Lowest price is not the same as best architecture.

One workload.
Five honest bills.