Last checked . Picks are based on the prices, rates, fees and terms each provider publishes, checked at the source. We never rank products by commission. See how we’re funded.
A runaway script, a leaked key or a loop in an agent can turn a small Gemini API bill into a large one overnight. Google now offers two ways to set a spend cap on that bill: a project-level cap inside Google AI Studio, and a spend cap budget in Google Cloud Billing. The AI Studio project caps were announced on March 16, 2026, and Google’s Cloud Billing page on spend cap budgets was last updated on Nov. 4, 2026. This guide covers how each spend cap works, the steps to set one, what happens when the limit is reached, and what Google’s pages leave unsaid.
Two kinds of spend cap for the Gemini API
The two tools overlap but are not the same. A project spend cap lives in AI Studio and limits Gemini API spending for one project. A spend cap budget lives in the Google Cloud console and limits one eligible service in one project. Google’s Cloud page lists the Gemini API, with the service name generativelanguage.googleapis.com, alongside Cloud Run, Cloud Run functions and the Gemini Enterprise Agent Platform, formerly Vertex AI.
A third limit applies whether or not anyone sets a spend cap. Google’s Gemini API billing page gives each usage tier a monthly ceiling for the whole billing account.
| Tier | How to qualify | Monthly billing account cap |
|---|---|---|
| Tier 1 | Set up and link an active billing account | $250 |
| Tier 2 | Pay $100, then wait 3 days from the first successful payment | $2,000 |
| Tier 3 | Pay $1,000, then wait 30 days from the first successful payment | $20,000 to $100,000 or more |
Google says spending is added up across every linked project that has the Gemini API turned on. When the total reaches the tier limit, service pauses for all of those projects until the next billing cycle starts on the 1st of the month. A custom spend cap is separate from this and can be set well below it.
How to set a spend cap in Google AI Studio
The billing page labels project spend caps as experimental. Accounts with the project editor, owner or admin role can set one. According to Google, the steps are:
- Open the Spend page in Google AI Studio, at aistudio.google.com/spend.
- Find the section named Monthly spend cap for the project.
- Select Edit spend cap and enter the dollar limit.
The announcement says the cap stays active until it is changed or switched off, and that it is “particularly useful” for accounts with several projects. Two caveats come from the billing page. Billing data can lag by about 10 minutes, so overages in that window are possible, and batch jobs and agent sessions may run past the cap. If a project moves to another billing account, its cap stays, but the spend counted against it resets to $0.
How to create a spend cap budget in Google Cloud
The Cloud version needs the Billing Account Administrator role, or the Billing Account Costs Manager role on the billing account plus the Editor role on each project. The permission behind it is billing.budgets.configureSpendCap. Google’s page lists these steps:
- In the Google Cloud console, open Budgets & alerts and select Create new budget.
- Under Define, choose Spend cap enforcement instead of Alerts only, enter a name and select Next.
- Under Scope, pick one project and one service, such as the Gemini API. The time range is fixed to monthly.
- Under Amount, enter the Target amount. The budget type is fixed to a specified amount.
- Under Actions, review the alerts. Google sends them to billing administrators and project owners at 50%, 80% and 100% of the target, and they cannot be customized. Select Finish.
An existing alerts-only budget cannot be converted. It has to be deleted and replaced with a new spend cap budget. Folders, organizations, labels and budgets that span several projects or services are not supported.
What happens when a spend cap is reached
Google’s Cloud page spells out what a spend cap does and does not stop.
| Item | What Google says |
|---|---|
| The capped service in the capped project | New usage is paused for the rest of the budget period |
| Other services and projects | Unaffected |
| Requests already in flight | They finish and still accrue charges |
| Persistent resources, such as storage | They keep running and keep accruing charges |
| Resources and data | Nothing is deleted |
| Reset | The cap resets at the start of the next budget period |
The Budgets page shows a Spend cap status column with four values.
| Status | Meaning |
|---|---|
| Not applicable | The budget is alerts-only |
| Configured | The cap is active but has not been triggered |
| Enforced | The cap was triggered and the service is paused |
| Lifted | The cap was enforced and then lifted by hand |
To lift an enforced cap, open Budgets & alerts, select View spend cap details in the banner, click the budget name, then choose Lift spend cap and Confirm. Google says the service can take up to an hour to return, and a lifted cap does not trigger again that month unless the target amount is raised. An enforced budget cannot be edited or deleted until it is lifted.
Prepaid billing as another limit
The billing page adds a different kind of limit: prepaid billing, which has been available since March 23, 2026. The rules it lists are below.
| Item | What Google says |
|---|---|
| Purchase size | $5 minimum, with $5,000 listed as the maximum |
| Expiry | Unused credits expire 12 months after purchase |
| Refunds | Generally not allowed, except when switching to postpay |
| Balance reaches $0 | Every API key in every project on the billing account stops working |
| Error at $0 | HTTP 402 Payment Required |
| Cutoff delay | Usage can run about 10 minutes past the balance |
| Auto-reload | Tops up when the balance falls below a trigger; the Monthly Auto-Charge Limit caps reloads per billing cycle |
| Switching plans | Postpay to prepay is available on the AI Studio Billing page; prepay to postpay is not supported |
Choosing a limit
These points are the Jolt Journal’s own suggestions, not Google’s. A Cloud spend cap budget gives a hard ceiling for the Gemini API, and Google’s own page advises setting the target slightly below the absolute limit, because enforcement is not instant and any overage is billed as normal. The AI Studio cap is quicker to set for a single project. Prepay suits anyone who wants spending to stop at a fixed pot of money.
Readers comparing what the models cost should start with the Jolt Journal’s ChatGPT, Claude and Gemini plan and API price comparison, and those who use the consumer app can see what changes on Nov. 13 in the report on Google AI Plus losing Gemini Pro. The guide to Gemini’s in-chat file editing covers a different part of the same product.
What the Gemini spend cap documentation leaves open
The AI Studio announcement does not say what a project spend cap does once it is reached, and the billing page does not name an error for it. It gives only the HTTP 402 error for an empty prepay balance. Google’s Cloud page does not give a general availability date for spend caps and refers to “this release,” so the list of eligible services may grow. The two pages also describe enforcement delay differently, at about 10 minutes for AI Studio and “not instant” for Cloud. Anyone relying on a cap to protect a budget should test it with a small limit first.
Editor’s take
The Gemini API now has several layers of protection, but none of them is a single switch. The tier cap applies automatically, the AI Studio cap is the quickest for one project, and the Cloud spend cap budget is the strictest option Google documents. Setting one of them takes a few minutes, and leaving all of them off means a leaked key can run up charges until the tier ceiling stops it.
This is general information about Google’s billing tools and not financial advice. Google’s documentation has the final word on limits, steps and prices.
