AI agents can work around the clock, but their bills can run just as freely. Google is giving businesses more ways to keep those costs under control.
Google introduced pay-as-you-go pricing, spending caps and new savings plans for Gemini Enterprise on Wednesday. The options are designed to help IT and finance teams manage unpredictable agent workloads without relying entirely on fixed per-user subscriptions.
For businesses experimenting with AI agents, the changes provide more control over how usage is funded, monitored and stopped before an unexpected workload turns into an unexpected invoice.
Gemini Enterprise adds pay-as-you-go pricing
Organizations can now combine Gemini Enterprise’s existing per-user subscriptions with a consumption-based option that charges according to agent and model usage.
The pay-as-you-go edition does not require an upfront commitment or base subscription fee. Charges vary with model and agent usage, allowing costs to rise or fall with demand instead of requiring businesses to pay for unused seats.
Google said the option is available to select customers and will roll out more broadly. The company did not provide a date for general availability.
Per-user subscriptions will remain available for organizations that want predictable monthly costs. Those subscriptions include daily quota pools shared across a Google Cloud project.
Businesses can also allow workloads to move automatically to pay-as-you-go billing after the pooled quota is exhausted. This could help teams avoid interruptions, although administrators would need to monitor the resulting overage charges.
The approach gives organizations a choice between fixed costs for employees who use AI consistently and consumption billing for developers or agents with less predictable workloads. It builds on Google’s broader effort to position Gemini Enterprise as a central platform for building and governing workplace AI agents.
Spending caps can pause AI agents
Google is also introducing project-level monthly spending caps through the Google Cloud Billing Console.
When a project reaches its limit, Gemini Enterprise agent activity will pause without affecting other workloads in the Google Cloud project, according to Google. Administrators can increase or remove the cap when they are ready for the affected agents to resume operating.
Automated emails notify administrators when a project reaches 50%, 80% and 100% of its budget. Google said its billing tools can also flag unusual spending patterns and identify the three stock-keeping units, or SKUs, contributing most to an increase.
A pricing calculator will allow teams to estimate costs across licenses, developer tools and agent runtimes before deploying a project.
These controls address a growing enterprise problem because AI agent workloads can vary substantially in token consumption, making their final cost difficult to predict before a task begins.
Google offers discounts for committed spending
Organizations with steadier workloads can use Gemini Enterprise Flexible Savings Plans to reduce token costs.
Google is offering a 10% discount for a one-year commitment and a 20% discount for a three-year commitment. Businesses select a monthly spending amount, with no stated minimum or maximum, and the committed spending can count against an existing Google Cloud enterprise agreement.
The savings plans are available to self-service customers and organizations with enterprise agreements, according to Google.
Google also plans to introduce deferred-execution pricing for certain workloads. Eligible tasks that do not need to run immediately could be scheduled during off-peak periods for discounts of up to 50%.
That option is not yet generally available. Google said it is coming soon for select workloads but did not announce a rollout date.
More Google coverage
- New Google Search AI Mode is ‘Total Reimagining,’ Says CEO Sundar Pichai
- In Major Ruling, Judge Finds Google ‘Willfully Acquired and Maintained Monopoly Power’ Over Digital Ad Market
- Google’s Big Bet on Nuclear Energy: ‘The Race to Power AI-Driven Data Centers is Accelerating’
- Computer History Museum Releases Original AlexNet Code: Why It Matters
Developer AI usage moves under one subscription
Access to Google Antigravity and AI features in Android Studio will also be included with Gemini Enterprise subscriptions for eligible customers.
Usage across Gemini Enterprise, Antigravity and supported developer tools will appear in a consolidated view rather than being divided among separate licenses and billing systems.
Daily allowances will be pooled across a project, meaning unused quota from business users can be applied to developer tools and custom agents. Google said access is available to select customers and will roll out more broadly.
The new choices represent a different approach from when Google incorporated Gemini into its Workspace subscriptions, which gave businesses a predictable per-user cost. Agent workloads can run for different lengths of time and consume varying numbers of tokens, making consumption harder to forecast.
IT and FinOps teams should establish project-level caps, decide whether overages will be permitted and monitor token use before scaling an agent beyond a limited deployment. They should also determine whether pausing an agent at its spending limit could interrupt a business-critical workflow.
Google’s changes make it easier to start using Gemini Enterprise without paying for unused seats. Whether they reduce AI spending will depend on how closely organizations monitor consumption once their agents begin working at scale.