State of the art of Serverless - Q1 2026 - Cost estimation #1

Go back to list

To ground the cost analysis in a realistic scenario, consider a document processing pipeline: a common enterprise workload where uploaded files are parsed, transformed, and stored asynchronously. Such a pipeline might handle invoice digitization, contract analysis, or log aggregation. By anchoring the benchmark to this archetype, it reflects a workload that is neither trivially lightweight nor computationally extreme, representative of the middle ground where the serverless cost model is most frequently debated.

Below, the base line used for estimate costs for running such workloads

Provider
AWS
Azure
Google
Oracle
T Cloud
Scaleway
Service Lambda x86 Lambda arm Container Apps * Cloud Run Functions Functions FunctionGraph✦ Serverless✦
Calculation
Requests $0.20 x 30 $0.56 x 30 $0.40 x 30 $0.0000002 x 30 $0.19 x 30 /
CPU / 0.000034 x 0.5 x 60 x 30000000 ✧ 0.0000336 x 60 x 30000000 / / 0.000011422 x 60 x 30000000
Memory 0.0000167 x 60 x 30000000 0.0000133 x 60 x 30000000 0.000004 x 60 x 30000000 0.0000035 x 60 x 30000000 0.00001417 x 60 x 30000000 0.000016667 x 60 x 30000000 0.0000011422 x 60 x 30000000
Estimation
Requests $6.00 $16.80 $12.00 $0.00 $5.70 /
CPU / $30,600 $60,480 / / $20,560
Memory $30,060 $23,940 $7,200 $6,300 $25,506 $30,000 $2,056
Total $30,066 $23,946 $37,817 $66,780 $25,506 $30,006 $22,616
  • *: Each container app manage 1 request maximum
  • ✧: Instances have a ratio of 1CPU:2GB, we selected 0.5CPU/1GB
  • ✦: From Euro during March 2026

Source