Why calculate cost per tenant instead of per user?

B2B SaaS usually prices per tenant (team, workspace, or company). Per-tenant economics show whether a fixed subscription price covers variable AI infrastructure and support costs.

What is included in the monthly per-tenant cost?

LLM API tokens, vector DB storage and queries, object storage, compute/containers, authentication, observability, support labor, and a share of fixed infrastructure.

How is gross margin calculated?

Gross margin = (monthly revenue - total monthly cost) / monthly revenue. It excludes one-time setup, sales, and marketing.

What does tenant break-even mean?

At the current price and cost structure, it is the number of active tenants needed for revenue to cover total monthly cost.

Does this include local GPU cost?

Compute infrastructure is entered as a monthly amount. For detailed GPU/local-vs-cloud sizing, pair this with the Inference Capacity Planner or GPU Cloud vs Buy Calculator.

🏢

AI Agent Multi-Tenant SaaS Cost Calculator

Estimate per-tenant and total monthly cost for a multi-tenant AI agent SaaS across LLM tokens, vector memory, storage, compute, auth, observability, support, and shared infrastructure.

Scenario presets

Tenant & usage profile

Total tenants Active tenant % Requests per active tenant / day Input tokens / request Output tokens / request Cached input (%) LLM provider Price per tenant / month ($) Projection months

Memory, storage & infrastructure

Vectors per tenant Vector DB Vector updates / tenant / month Vector queries / tenant / month Storage per tenant (GB) Object store Storage reads / tenant / month Storage writes / tenant / month Compute / containers monthly ($) Auth provider Observability tool Shared infra monthly ($) Support hours / 100 active tenants / month Support hourly rate ($)

Estimated cost & margin

Active tenants

Monthly requests

Cost per tenant / month

Total monthly cost

Monthly revenue

Gross margin

Tenant break-even

Period cost

Cost breakdown

Component	Monthly	% of total
Enter inputs to see breakdown.

Verdict