① Google achieved a significant leap in artificial intelligence in 2025. At the beginning of the year, it was still considered by Wall Street to be lagging behind OpenAI, but now it has forced OpenAI to sound a “red alert”; ② Google's success can be summarized in three aspects: its accumulated technical strength, a well-established ecosystem, and correct decisions made by leadership, which might also provide development directions for other AI companies.
$Alphabet-C (GOOG.US)$ In 2025, the company achieved a significant breakthrough in artificial intelligence. At the beginning of the year, it was widely believed by Wall Street to be lagging far behind OpenAI in the AI race, but this year it managed to trigger a 'red alert' within OpenAI.
Gene Munster, managing partner at Deepwater Asset Management, noted in a report that Google is arguably the strongest in terms of a fully integrated AI technology stack. Moreover, the user base of the Gemini model is growing faster than OpenAI’s ChatGPT, and Google Cloud continues to maintain its leading position during the infrastructure construction cycle.
The multi-faceted progress across Google’s operations indicates strong performance results for this year and lays a foundation for success in the coming year. Given that all AI competitors have been relentless during the same period, this raises curiosity about how Google managed to reverse its fortunes within just one year.
Technological Strength
Over the past two years, Google has faced skepticism regarding the inefficiency often associated with large corporations. However, the tech giant did not lose heart but instead actively integrated resources, leveraging collaboration across its various business lines and its economies of scale to achieve a “latecomer advantage.”
As early as April 2023, Google merged Google Brain and DeepMind, two top-tier research teams, to create a cutting-edge model. These two teams had previously introduced the Transformer AI architecture and the AI Go champion AlphaGo, respectively, both demonstrating indisputable strength in the technical field.
Close cooperation between these two teams was one of the main reasons for Gemini 3’s breakthrough. This model saw simultaneous advancements in reasoning, multimodal capabilities, and programming, making it the first model from Google considered capable of posing a substantive threat to OpenAI.
Even so, Gemini 3 cannot yet be called a transformative model, and Google’s ambitions are not limited to foundational models alone. As a tech giant, Google’s advantage also lies in its more mature and comprehensive ecosystem.
Demis Hassabis, CEO of DeepMind, once pointed out that Google is integrating Gemini into products such as Maps, YouTube, Android, Search, and Workspace. This distribution network and end-user data feedback loop constitute an insurmountable moat. This statement also reflects that Google is shifting its focus toward practical application scenarios.
Ecological moat
The breakthrough of Gemini 3 marks Google's return to the global top tier in foundational models, but this is only one aspect of Wall Street's optimism about Google. More importantly, Google has presented the prospects of implementing and commercializing artificial intelligence to investors.
Needham analyst Laura Martin emphasized at Gemini that Google's business units, such as Search and YouTube, have strategic advantages in distributing AI products, allowing them to quickly achieve a return on investment.
Starting in May, Google fully introduced the agent model into its ecosystem and demonstrated the ability to automatically execute complex tasks across applications. When Gemini 3 was released in November, Google integrated the new model into its search products for the first time while also making it available to developers, thereby creating a leading market implementation pace.
By integrating the model into its existing business, Google has easily created a window to reach billions of users for its artificial intelligence business, helping to achieve rapid iteration and enhance profitability through continuous testing.
At the same time, another star business, Google Cloud, has also brought substantial returns to Google. In the third quarter, Google Cloud's revenue grew by 33.5% year-on-year, with an operating profit of up to 23.7%. Google's custom chip TPU has reduced Google's external dependencies and enhanced the cost advantages of large models, thereby solidifying the financial foundation for its artificial intelligence business.
This set of TPU-Gemini scenario combinations has created an effective investment-return-reinvestment method for Google, significantly shortening the company's product cycle from research and development to implementation. This system is the key to its clear advantage over OpenAI and its leadership in the market.
Talent advantage
On Google's path to technological resurgence this year, another person has been frequently mentioned: Josh Woodward, who is in charge of the Gemini application. This rising star at Google has become one of the hottest role models in Silicon Valley thanks to the breakout success of Gemini 3.
Woodward led Google's major efforts to arm its ecosystem with AI, while also serving as the head of Google Labs, the home of Google's experimental artificial intelligence projects. Before his promotion in April, Google's position in the field of artificial intelligence was quite awkward, with its parent company Alphabet's stock price plummeting 18% in the first quarter amid market skepticism.
August marked the first turning point for Google. The Woodward team launched an image generator called Nano Banana, a feature of Gemini that allows users to blend multiple photos together to create personalized digital characters.
This model gained immense popularity immediately after its launch. By the end of September, Gemini had processed over 5 billion images, prompting internal comments at Google describing it as "so popular it melted the TPUs."
The application also significantly increased Gemini's active user base. In October, Google revealed that the monthly active users of the Gemini app surged from 350 million in March to 650 million. For reference, OpenAI also announced in October that ChatGPT reached 800 million weekly active users.
Clay Bavor, co-founder of the AI startup Sierra, stated that Woodward was one of the earliest at Google to recognize the potential of large language models in product development. He possesses the ability to fully understand a new technology, foresee future directions, and identify its evolution and usage patterns.
This highlights the critical role of human ingenuity in the advancement of artificial intelligence. As the industry explores ways to move forward rapidly, individuals with clear goals and direction demonstrate their invaluable worth. By securing this talent, Google achieved a remarkable turnaround.
Editor/Doris