Getting Started with Snowflake Per-User Quotas for User-Level Cost Management of Warehouses and AI Features

#snowflake #ai #data

Note:
This article represents my personal views and not those of Snowflake.

Introduction

Snowflake's Per-User Quota has reached Public Preview (PuPr). Per-User Quota is a feature that lets you set per-user credit limits, helping administrators govern how individual users consume Snowflake resources. In this article, we'll explore the feature's concepts and how to think about adoption.

Note:
As of August 3, 2026, this feature is in Public Preview and may be subject to changes in the future.

Background: Why Per-User Quotas?

Recent Snowflake releases have introduced more services that are consumed at the individual user level — not just AI Functions, but also Snowflake CoCo, Cortex Agents, and Snowflake CoWork. This shift has introduced challenges such as: a small number of users consuming a disproportionate share of warehouse or AI credits, delayed awareness of cost overruns, the need for manual admin intervention when limits are breached, and difficulty tracing who is consuming how much.
Per-User Quota addresses these challenges by letting you enforce per-user credit caps, and optionally block further usage automatically when those caps are hit.

What Can You Do with Per-User Quota?

Per-User Quota enables the following capabilities:

Set credit limits per user on a defined cycle
- Warehouses support monthly limits only; AI-related limits can be set monthly, daily, or both
- Monthly and daily AI limits are evaluated independently and both reset on a UTC monthly/daily boundary. Custom cycles (e.g., weekly) are not currently supported
Target all users, or a subset defined by tags
- You can target all users while excluding specific tagged users, or reassign users to different quotas by changing their tags
Take action when a limit is reached
- For warehouses, you can invoke a custom action (any stored procedure)
- For AI-related requests, a built-in block mechanism is available (note: this block feature is not available for warehouses)
All users in a single quota share the same limit — per-user overrides within a quota are not supported
- If you need different limits for different users, create separate quotas and assign users via tags. For example, you could create a "Standard" quota and a "Power User" quota, and tag users accordingly. See Respond to escalations in the documentation for examples

What Resources Can Be Targeted?

The resources that Per-User Quota can manage are listed below. In the documentation, these are referred to as "domains." When configuring a Per-User Quota, you specify one or more domains to monitor.

Domain	What's Monitored	Cycle Options
Warehouse	Credit usage for compute resources powering query execution, including warehouse metering and query acceleration costs	Monthly only
AI Functions	Credit usage for AI function operations attributed to the user	Monthly, daily, or both
Snowflake Intelligence	Credit usage for Snowflake CoWork (formerly Snowflake Intelligence) operations	Monthly, daily, or both
Cortex Agents	Credit usage for Cortex Agents operations attributed to the user	Monthly, daily, or both
CoCo	Credit usage for Snowflake CoCo (formerly Cortex Code) operations attributed to the user, including spend from CoCo CLI, CoCo in Snowsight, and CoCo Desktop	Monthly, daily, or both

Important
Within a single quota, you can monitor either the Warehouse domain or AI domains — but not both together. This is because warehouses are measured in credits while AI domains are measured in AI credits, and the two units cannot be mixed.

How Does This Differ from CoCo's Built-in Limits?

CoCo has its own dedicated parameters for setting daily credit usage limits per user, separately for CLI, Desktop, and Snowsight (see Cost controls for CoCo). These are evaluated daily and can be configured at both the account level and the user level, with user-level settings taking precedence.
Here's how Per-User Quota compares to those CoCo-specific parameters:

Dimension	Per-User Quota	Cortex Code Parameters
Scope	Spans multiple AI domains	CoCo only
Cycle	Monthly and/or daily	Daily only
Targeting	All users, or users filtered by tags	Account-level or user-level only
Monitoring	Integrated with Snowsight's budget management; supports notifications and limit-triggered actions	Usage visibility via views; can suspend usage

If you want to quickly control CoCo costs alone, or apply a uniform account-wide cap, the CoCo-specific parameters are the simpler path. If you want cross-cutting control across warehouses and AI features, Per-User Quota is the right tool.
Per-User Quota also stands out for its flexible targeting via tags and its integration into Snowsight's broader budget management capabilities.

What Problems Does Per-User Quota Solve?

Earlier, we identified these challenges:

For example: a small number of users consuming a disproportionate share of warehouse or AI credits, delayed awareness of cost overruns, the need for manual admin intervention when limits are breached, and difficulty tracing who is consuming how much.

Here's how Per-User Quota addresses each of them:

Challenge	How Per-User Quota Helps
A small number of users consuming excessive warehouse or AI credits	Set per-user limits. Warehouses support monthly caps; AI features support monthly and/or daily caps
Delayed awareness of cost overruns	Email notifications to the user and admin when usage reaches a configured threshold percentage
Manual admin intervention required when limits are exceeded	Automatic custom action (for warehouses) or built-in block (for AI features); automatic reset at the start of each new quota cycle
Difficulty tracing per-user consumption	Centralized visibility into all users' usage via the Snowsight dashboard

Adoption Guidance

A phased rollout is recommended to minimize disruption. The general pattern is: always enable notifications, and decide per-phase whether to activate custom actions or automatic blocking.

Phase	Recommended Configuration
Internal AI PoC / Dev/Test	Set conservative monthly limits with notifications only; do not enable auto-blocking for AI features
Org-wide rollout / Production	Use tags to segment by department or role; enable custom actions for warehouses and auto-blocking for AI features
Governance hardening	Configure multi-threshold alerts at 50% / 80% / 100%; both users and admins receive notifications

Conclusion

In this article, we explored Per-User Quota, which recently entered Public Preview. Snowflake continues to evolve rapidly, and user-facing services like CoWork and CoCo are driving more per-user consumption patterns than ever. Alongside that growth, cost governance at the user level becomes increasingly important. Per-User Quota provides a solid foundation: per-user credit caps, threshold-based notifications, enforcement mechanisms, and centralized monitoring integrated into Snowsight. We hope you'll give it a try when it reaches GA and make it part of your cost governance strategy.