AI Gateway Metering and Licensing
AI Gateway uses a consumption-based token metering model funded through Software NGFW credits, with tiered volume discounts.
| Where Can I Use This? | What Do I Need? |
|---|
AI Gateway employs a consumption-based metering model. All AI traffic — including
MCP, LLM, and A2A payloads routed through the Gateway — is measured in
tokens.
Token consumption calculation:
- Prompts (request/responses) — token values returned by the LLM.
- MCP tool calls — industry-standard conversion rate of 1 token per 4 plaintext
(UTF-8) characters.
- A2A requests — industry-standard conversion rate of 1 token per 4 plaintext
(UTF-8) characters.
This standard applies consistently across both SaaS and Hybrid deployment
environments.
In alignment with the broader Prisma AIRS portfolio, AI Gateway uses
Software NGFW credits for
licensing. Total credit requirements are estimated based on your anticipated token
usage. The model includes built-in tiered volume discounts — as you reach higher
usage tiers, the cost per billion tokens decreases.
Contact your Palo Alto Networks sales representative with your expected token
consumption, and your account team will calculate the exact credits needed for your
deployment.
