Google is changing its billing and resource allocation for its AI-driven developer platforms -- specifically Google Antigravity in
Gemini Enterprise -- to help brands and agencies of all sizes control rising costs.
On Wednesday, the company said it will offer businesses ways to manage AI costs through flexible
billing options.
It provides cost controls for AI agents and subscriptions for developer platforms such as Antigravity and Android Studio AI in Gemini Enterprise.
Autonomous AI agents consume much more compute power than standard AI chatbots, which creates severe and unpredictable budget risks for businesses including agencies and
brands.
advertisement
advertisement
Unlike a standard "one-request, one-response" chatbot, Antigravity 2.0 acts as an autonomous command center where multiple background agents execute multiple
loops that run tests, browse code bases, and bring on subagents entirely on their own -- sometimes with little control, which can cause costs to skyrocket.
The new payment options include a
flexible savings plan when companies commit to a set monthly spend if their AI workloads are steady or increasing.
Businesses also can set monthly caps on AI spend, the company wrote in a blog
post published Wednesday.
If AI workloads remain steady or rise, the “Flexible Savings Plans” allows the company to commit to a monthly spend, which enables businesses to reduce
token costs by between 10% and 20% with no minimums, no maximums, and no new billing silo to manage.
Industry executives and brands have noted in the past that AI costs are expensive, and
many have not determined how to prove the return on investments. Without that ability, many have not been able to get the signoff from upper management in the company.
For large enterprises
generating revenue of $1 billion or more, the average spend on AI licenses and infrastructure is about $118,000 per month, according to a 2026 benchmark report from Eliassen Group, a strategic
consulting company.
Gartner has published yearly global forecasts and year-over-year increases. The analyst firm projects worldwide AI spending about $1.765 trillion in 2025, growing to
about $2.596 trillion in 2026.
For companies using the Gemini Enterprise app per-user seat subscription, for example, businesses can pay a fixed monthly fee per user,
which includes daily quota pools that are shared across the entire project.
Google said it can optimize spend through predictable budgeting, providing finance teams with
a monthly baseline for teams with consistent daily needs.
Google also will soon offer a pay-as-you-go option for some customers using the Gemini Enterprise
app.
There is no upfront commitment or base subscription fee, and businesses only pay for the compute and tokens that the teams consume at standard model API rates.
The spend scales up and down automatically as the system is used.
Google Antigravity in Gemini Enterprise, an agent-first developer platform that brings agentic
coding and agent-building capabilities to technical teams, also will roll out soon.