Google’s Gemini hits 1 billion users as cash flow turns negative amid AI spending

Google CEO Sundar Pichai has announced that the company’s Gemini artificial intelligence product has reached 1 billion monthly active users, establishing it as the fastest-growing product in the firm’s history. The milestone makes Gemini the 14th Google product to achieve this scale, placing it on par with OpenAI’s ChatGPT, which surpassed the 1 billion user mark in June or August depending on the source.

The metric specifically counts users who actively open the Gemini app or visit the web interface to enter a prompt or access Gemini Live. While Gemini is integrated into Gmail, Drive, and Google Search, these integrations are not counted in the 1 billion MAU figure. Google’s vice-president Josh Woodward noted that more than 100 million of these users are on iOS, requiring manual downloads rather than the pre-installation seen on Android devices.

Usage data reveals significant engagement patterns, with 63 per cent of active users utilising voice input and an increasing number adopting voice-only interactions. Of those using Gemini Live, 20 per cent share camera feeds or screens with the assistant. The platform also generates 150 million images daily, all watermarked with SynthID, while 38 per cent of school-related requests include an attachment, prompting Google to develop new study tools.

Despite the user surge, Google faces significant operational headwinds. The company’s spending on AI infrastructure has pushed its cash flow into negative territory for the first time ever. Google does not break out AI-specific revenue in its earnings reports, leaving the financial return on this heavy investment unclear to investors.

Internal challenges are also mounting. DeepMind co-founder Demis Hassabis has stepped down from his leadership role, though he remains at the company; reports suggest he is not interested in Google’s hyperscaling AI route. Additionally, Google has lost several key scientists and researchers in recent months, contributing to delays in the release of the Gemini 3.5 Pro model, which was promised for June.

Reports indicate that Google is unhappy with Gemini Pro’s coding performance, which lags behind competitors such as OpenAI and Anthropic. The company has begun training Gemini 4 while still working on the delayed 3.5 Pro, but it remains unclear if the latter will ever be released. The cracks in the product roadmap are beginning to show, raising questions about whether dominant market presence can sustain user numbers if the underlying technology underperforms.