How Google put together the pieces for its AI comeback

In 2022, the introduction of ChatGPT caught Google off guard, but recent developments have sparked a resurgence of interest in the tech giant's artificial intelligence initiatives. This November, Google unveiled the Ironwood AI chip—its latest tensor processing unit (TPU)—which the company claims enables users to operate and expand some of the most data-intensive models available today. Alongside Ironwood, Google introduced Gemini 3, its newest AI model designed to require less prompting while delivering more intelligent responses. Salesforce CEO Marc Benioff, who has utilized OpenAI's ChatGPT for three years, expressed his enthusiasm for Gemini 3 in a post on X, stating that after just two hours of use, he felt the model was a game-changer. "The leap is insane," he remarked, noting the enhanced sharpness and speed of the AI's capabilities. As Alphabet's stock surged over 5% on Monday, following a more than 8% rise the previous week, it marked a clear shift in investor sentiment. Warren Buffett's Berkshire Hathaway disclosed a $4.3 billion stake in Alphabet, contributing to a nearly 70% increase in Alphabet shares this year, significantly outperforming Meta. The company's market capitalization also recently eclipsed that of Microsoft. Despite Nvidia's impressive third-quarter earnings, which exceeded expectations, analysts have pointed to Google's developments as a cause for concern among AI competitors. According to Melius Research analyst Ben Reitzes, the resurgence of Alphabet in the AI space is prompting a reevaluation of the competitive landscape. Experts agree that while Google has regained momentum, its position remains precarious in the highly competitive AI market. Michael Nathanson, co-founder of Moffett Nathanson, noted that Google CEO Sundar Pichai appears to have successfully aligned the company's AI offerings, addressing earlier criticisms of stagnation. Google's previous missteps, such as the temporary withdrawal of Imagen 2 due to inaccuracies, highlight the challenges the company faced in catching up with OpenAI. However, the swift rollout of Gemini 3 after the successful Gemini 2.5 demonstrates Google's agility in the space. The recent launch of Nano Banana, an advanced image generation tool, has further solidified Google's standing, as it quickly rose to the top of the Apple App Store, surpassing ChatGPT. Coupled with its vast resources from YouTube, Google holds a significant advantage in training models for both image and video generation. Moreover, Google’s cloud services have seen substantial growth, marking its first $100 billion quarter, bolstered by AI applications. The Ironwood chip, noted for its efficiency—reportedly 30 times more power-efficient than Google's first TPU—positions the company as a formidable player in the AI chip market. While Google has made significant strides, experts caution that the race is far from over. The competition remains fierce, with rivals like Anthropic and OpenAI also advancing their models. Analysts suggest that the future will hinge on which companies are willing to invest heavily in their AI capabilities, as the landscape evolves. Google executives have acknowledged the necessity of doubling their serving capacity every six months to meet the growing demand for AI services. Despite its recent successes, the company still faces challenges in maintaining user engagement and addressing issues with its consumer chat product, which lags behind OpenAI's offerings in terms of user numbers. In summary, while Google's recent advancements in AI have rekindled optimism about its prospects, the journey ahead is fraught with challenges as it competes in a rapidly changing environment.