Gemini 3 drives significant business growth, Google AI model applications double in five months.

```

As model quality improves, sales of Google’s Gemini AI model have seen explosive growth in the past year.

On January 19, local time, The Information cited Google’s internal data showing that requests for the Gemini API (Application Programming Interface) via Google Cloud are surging rapidly. Data shows that API call volume jumped from about 35 billion when Gemini 2.5 was released last March, to 85 billion in August, more than doubling.

Sources revealed that the release of Gemini 3 last November triggered another wave of usage and received widespread praise.

This growth is not only reflected in quantity, but also in improvements in quality and profit margin. It is reported that the early Gemini 1.0 and 1.5 models had negative profit margins due to steep discounts. However, with the launch of Gemini 2.5 and subsequent versions, improvements in model quality have allowed Google to shift from a pure “price war” to a “quality war”, resulting in positive marginal profits.

Market Test under Massive Capital Expenditure

Although business data is positive, the market still focuses on the high input-output ratio.

Last fall, Google estimated that its capital expenditure (Capex) would be between $91 billion and $93 billion, nearly double the $52.5 billion in spending planned for 2024.

Investors are closely watching the upcoming Q4 earnings report for signs that these massive investments are paying off.

Opportunities and Challenges for Enterprise Applications

On the software application front, Google is trying to boost profit margins through Gemini Enterprise. A Google spokesperson revealed that Gemini Enterprise currently has 8 million subscribers from 1,500 companies, in addition to over 1 million online registered users.

The spokesperson said: "We’re seeing tremendous momentum across the cloud business, especially in our AI applications."

However, market feedback is polarized. Simon Margolis from consulting firm Sada pointed out that client feedback is not entirely positive. He said bluntly, "The ratio of customers who like it vs. dislike it is about 50/50."

Some challenges arise from Google’s “developer-first” DNA. Margolis analyzed: "Google has always been more like a ‘build-your-own’ cloud rather than a ‘buy-a-product’ cloud." This means many clients prefer to use Gemini models directly to build their own custom agents, instead of buying Google’s pre-made software suites.

Chirag Mehta, chief analyst at Constellation Research, also noted that while Gemini Enterprise performs well in answering general questions based on enterprise data, it still struggles with specific tasks. However, he emphasized that clients are not abandoning the product as they have with some competitors, but are continuing to use it with a "let’s give it a try" attitude.

Risk Disclaimer and TermsThe market has risks; investments should be made cautiously. This article does not constitute personal investment advice, nor does it take into account the particular investment objectives, financial situation, or needs of any individual user. Users should consider whether any opinions, viewpoints, or conclusions in this article are suitable for their specific circumstances. Investing based on this article is at your own risk.

```