Google Adds Antigravity to Gemini Enterprise With New AI Spending Controls

Google is expanding the enterprise version of its Antigravity coding-agent platform by bringing it under Gemini Enterprise and adding centralized controls for budgets, usage, and overages. The update targets a familiar problem for companies adopting AI coding tools: developers may gain productivity, but token consumption and cloud spending can become difficult to predict.cloud.google+1

Under the new approach, organizations can manage Antigravity, Gemini Enterprise, Android Studio AI features, and custom agent workloads through a more unified billing and governance model.

From Separate Usage to Shared Quotas

Previously, Antigravity usage was primarily tied to individual cloud projects and consumption-based billing. That gave companies access to Google Cloud’s existing security and governance features, but it made it harder to manage usage across large engineering organizations.

Google is now pooling the developer-tool quota included in eligible Gemini Enterprise subscriptions at the Google Cloud project level. Instead of assigning a fixed allowance to each developer, teams can draw from one shared capacity pool. Unused quota from employees with light usage can therefore support colleagues or teams with higher demand.cloud.google+1

This model can make licensing more efficient, particularly in organizations where some developers use AI coding tools daily while others only use them occasionally.

Budget Caps and Overage Controls

The update gives administrators several new mechanisms for limiting unexpected costs:

• Monthly project-level spending caps can be set through Google Cloud Billing. When a project reaches its limit, eligible agent API requests are paused to prevent further charges.
• Budget alerts can notify administrators when spending reaches 50%, 80%, and 100% of the defined budget.
• Overage enablement allows workloads to continue after the pooled quota is exhausted, using pay-as-you-go pricing while remaining subject to the monthly spending limit.
• Centralized usage metrics show quota consumption, pay-as-you-go usage, API activity, and 30-day billing trends.cloud.google+1

For example, a company could allocate a monthly budget to an internal development project. If an Antigravity agent begins making unusually large numbers of API calls, administrators can see the rise in consumption, identify the relevant workload, and decide whether to pause it or allow controlled overage spending.

Better Visibility for FinOps Teams

The main benefit is not simply a lower bill. It is clearer accountability.

In many cloud environments, finance teams can see the total cost of a project but cannot easily determine which development team, workflow, or agent generated that expense. Google’s new usage tools aim to provide more detailed visibility into token consumption, API calls, and developer activity.cloud.google

Google is also introducing anomaly detection for unusual AI spending. The system can flag unexpected cost increases and identify the top billing items driving the change, helping FinOps and engineering teams investigate before a spike turns into a major monthly overrun.itbrief.co+1

Subscription and Consumption Options

Google is keeping subscription-based pricing while adding more flexible consumption options. Companies with steady usage can retain per-user Gemini Enterprise subscriptions, while organizations with variable demand can use pay-as-you-go billing based on token and compute consumption.itbrief.co+1

For customers with predictable workloads, Google also offers Flexible Savings Plans. The company says these plans can reduce token costs by 10% for a one-year commitment or 20% for a three-year commitment.ciodive+1

Antigravity access through Gemini Enterprise is initially available to select customers, with wider availability planned. Some organizations may therefore need additional eligible Gemini Enterprise licenses before they can use pooled quota across their developer tools.investing+1

Google’s update reflects a broader shift in enterprise AI: coding agents are no longer treated as small productivity add-ons. They are becoming operational workloads with budgets, limits, monitoring requirements, and financial governance comparable to other cloud infrastructure