2024 is coming to an end, and the AI war between Google and OpenAI is in full swing. Google CEO Pichai’s annual strategy conference speech indicates that 2025 will become a key turning point in the field of AI. Faced with the strong rise of OpenAI and the huge influence of ChatGPT, Google is actively fighting back and trying to regain its AI leadership position with its Gemini series of products. This speech is not only a summary of the competition in the past two years, but also a clear plan for future development. How Google will respond to challenges and whether it can successfully counterattack is worthy of attention.
2025 has not yet arrived, and the AI war in Silicon Valley has already been ignited. Google CEO Pichai issued an inspiring call to all employees at the annual strategy meeting: 2025 will be a crucial year! He emphasized that Google must move forward at an unprecedented speed, because this is not only a technology competition, but also a technological competition. A disruptive moment related to the future AI landscape.
This statement is undoubtedly a response to the fierce competition between Google and OpenAI. Since the advent of ChatGPT, Google’s leadership in the field of AI has been challenged unprecedentedly. In the past two years, OpenAI has overshadowed almost all of Google’s efforts. However, in December 2023, Google made a strong counterattack with a series of blockbuster products such as Gemini2.0 Flash and Veo2, showing the potential to surpass the leading model. The turning point in this AI competition seems to have arrived.
At the meeting, Pichai showed a large model comparison chart. The performance of Gemini 1.5 was clearly ahead of competitors including OpenAI. When asked about how ChatGPT has become synonymous with AI, Pichai said that Google does not always have to be the first, but it must be the best in its category. The significance of 2025 is that Google must consolidate its first-tier position in the field of AI. Logan Kilpatrick, head of Google developer relations, has also hinted many times that Google will make full efforts next year and is expected to officially launch the full version of Gemini 2.0 in January. Hassabis, CEO of Google DeepMind, even bluntly said that this is just the appetizer, and the fun will begin in 2025.
Some analysts predict that Google may open more powerful models for free next year and release competitive models that directly challenge GPT-3. In order to occupy the commanding heights in this AI revolution, Google is going all out.
Gemini: Google’s bet on the next 500 million apps
For Google, a more serious challenge comes from OpenAI's cannibalization of the search business. While Google still dominates the search market, generative AI is giving people new ways to access information online. OpenAI is being regarded by more and more people as synonymous with AI, just like Google's position in the search field.
Google is under tremendous pressure and has invested heavily in developing Gemini to consolidate its lead in the AI field. The Gemini app allows users to access a variety of tools including chatbots. Pichai emphasized that "building large new businesses" is a top priority. Google currently has 15 apps with more than 500 million users, and executives generally believe Gemini will be next. While Pichai sees "strong momentum" for Gemini applications, he acknowledges that there is still much work to be done in 2025 to close the gap and establish leadership. "The biggest focus next year will be expanding Gemini on the consumer side," he added.
In addition, Google faces multiple legal actions in the United States, including accusations of monopolizing search and illegally dominating online advertising technology. British regulators also initially determined that Google’s advertising technology practices affected the country’s competitiveness. In this regard, Pichai said that this is a challenge brought by Google's scale and success, and it must ensure that it is not distracted.
Google: Can "latecomers" catch up?
Data from OpenRouterAI shows that Google Gemini's market share among developers has soared from about 5% in September to more than 50%, far ahead, and drove the stock price up 14%. At this time, it has been two years since OpenAI released ChatGPT.
As early as 2017, Sam Altman regarded Google DeepMind as the most powerful opponent in the AGI competition. However, Google, once recognized as an AI giant, appeared to be somewhat "powerless" after the release of ChatGPT. Google not only made frequent mistakes in product launches, but its stock price also fell.
What is puzzling is that Google has the best hardware, the most training data and the best talents in the world. Why is it lagging behind OpenAI in the AI race, and even startups like Anthropic have difficulty matching it? Some analysts believe that, Google's "ship is about to turn around". Due to security considerations, Google needs more time to conduct compliance inspections. In early 2023, Google released Bard in a hurry, but its stock price plummeted due to wrong answers. In addition, as a large company, Google is also subject to more legal restrictions, such as copyright issues on user data.
Nonetheless, Google is re-emerging as king. Its future goal is to continue to lead the development of "intelligent agents" and integrate them into browsers and mobile phones to automate user tasks.
Google’s big AI events in 2024
2024 is the year when Google AI will “strengthen”. Many well-known products were released this year, such as various versions of Gemini, NotebookLM and the AI capabilities of Pixel phones.
Google claims to have “60 major AI releases” in 2024, including:
Gemini model: Google launched its first native multi-modal model Gemini1.0 in December last year, and has since launched Gemini1.5 and Gemini2.0 Flash versions. These models have significantly improved in terms of multi-modal processing capabilities, context windows, and reasoning performance. Gemini2.0 integrates multi-modal and native tool usage capabilities, marking the official entry of large models into the "intelligent agent" era.
NotebookLM: This AI-driven scientific research and writing assistant can better help users understand and explore complex materials after the underlying model is switched to Gemini1.5Pro. Its key feature, Audio Overview, can also generate audio of two AI hosts discussing each other, assisting users in summarizing and in-depth discussions of materials.
AlphaFold3: Google's AlphaFold3 has improved its performance by at least 50% in predicting interactions between proteins and other molecular types, bringing more possibilities to the field of biological sciences.
Pixel mobile terminal: Google provides a series of AI functions based on Gemini and Imagen2 for the new Galaxy S24 series and Pixel9 mobile phones, including summary capabilities, generative photo editing and travel planning, etc., aiming to provide users with smarter and more convenient features through AI. experience.
Google is going all out and looks forward to bringing more surprises to users next year.
All in all, Google’s competition in the AI field has not been smooth sailing, but its strong technical strength and resource accumulation have given it strong competitiveness. In 2025, let us wait and see whether Google can successfully achieve overtake with its Gemini series of products and other innovative achievements, and who will win in the end.