Google is seeing results not only in expanding chatbot users but also in sales of application programming interfaces (APIs) to external companies and developers, backed by its latest Gemini AI model with significantly improved performance.

Sales of APIs that allow access to Gemini AI models have surged over the past year, The Information reported on Jan. 19 local time, citing three sources familiar with the matter.

That likely also led to higher revenue from sales of the server infrastructure that Google Cloud provides for AI.

Google provides access to Gemini models through APIs available on Google Cloud. API calls, meaning requests sent to the Gemini API, increased to 85 billion in August from 35 billion in March, when the Gemini 2.5 model was first unveiled, The Information reported.

Given the positive response to the Gemini 3 series released in November, API revenue may have increased further in the second half of the year.

That performance is expected to show up in numbers in Google’s fourth-quarter earnings release on Feb. 4, The Information reported.

Investors have been watching whether the company can show results in numbers after investing heavily in building AI infrastructure. Google last fall estimated its 2025 capital spending would reach $91 billion to $93 billion. That is nearly double $52.5 billion in 2024.

A major challenge Google currently faces is persuading companies to pay for and use software developed on the basis of Gemini models, The Information reported.

A key example is Gemini Enterprise, which combines a Gemini chatbot with enterprise data search capabilities with a platform that can build and use AI agents. Google will also highlight Gemini Enterprise growth in its February earnings release, The Information reported.