In the just-released Q4 2025 earnings report, Alphabet (Google's parent company) delivered a market-boosting performance: its AI chatbot Gemini has officially surpassed 750 million monthly active users (MAU). This significant growth not only marks the successful consumer adoption of the Gemini3 model, but also helped Alphabet break through the $40 billion revenue threshold for the first time in its history.
Driving Force: Deep Transformation of Gemini3
Google CEO Sundar Pichai emphasized at the earnings call that the release of Gemini3 was the key turning point for the company this quarter. As Google's most advanced model to date, Gemini3's significant improvements in reasoning depth and multimodal interaction directly boosted user engagement:
Explosive User Growth: Monthly active users jumped from 650 million last quarter to 750 million, firmly placing it among the top tier of the AI industry.
Scalability Efficiency: Pichai revealed that the Gemini API currently processes over 100 billion tokens per minute, and with model optimization, its service cost per unit dropped by 78% in 2025.
Market Position: Catching Up with ChatGPT, Leading Meta AI
Although Gemini is showing strong growth, the competitive landscape remains tight. The current global AI monthly active user rankings show a three-way standoff:
ChatGPT: Still leading the pack, with an estimated 810 million monthly active users by the end of 2025.
Gemini: At 750 million, it follows closely, with the gap rapidly narrowing.
Meta AI: At nearly 500 million, it ranks third.
To further penetrate the market, Google recently launched a monthly $7.99 Google AI Plus package. Although its financial contribution hasn't been reflected in the fourth quarter, analysts believe that this more affordable pricing strategy will be a powerful tool for attracting budget-sensitive users in 2026.
Hardware Breakthrough: Ironwood TPU Challenges NVIDIA
Beyond the application side, Google's layout in AI infrastructure is also aggressive. The report disclosed that Google's latest generation TPU acceleration chip Ironwood (TPU v7) has been officially released.
Performance Comparison: Ironwood's FP8 performance reaches 4.6 petaFLOPS, comparable to NVIDIA's B200.
Scale Advantage: Ironwood supports integrating more than 9216 TPUs in a single superpod, aiming to reduce AI computing costs through a vertically integrated software and hardware ecosystem, directly challenging NVIDIA's dominance in computing power.
Chief Commercial Officer Philipp Schindler summarized that whether it's expanding free-tier users or growing Gemini Enterprise subscriptions, AI is driving business growth across the board. Alphabet has announced that its 2026 capital expenditure guidance will reach between $175 billion and $185 billion
