Google CEO Sundar Pichai has urged his AI team to take a break after a grueling sprint to launch Gemini 3, saying the group deserves some rest. Speaking on the "Google AI: Release Notes" podcast released Wednesday, Pichai acknowledged the intense effort behind the latest AI model, joking, “I think some folks need some sleep,” and expressing hope that he and his team can “get a bit of rest.” The launch of Gemini 3 on November 18 marked a major milestone for Google, coming amid a surge in the company’s market value. Google’s stock has climbed nearly 70% this year, including a 12% jump immediately following the Gemini 3 release, pushing the company’s market capitalization close to $4 trillion. Gemini 3 has been widely praised for its advancements in reasoning, speed, and multimodal capabilities. Salesforce CEO Marc Benioff hailed the model as a transformative leap, stating on X that after just two hours of use, he’s “not going back” to ChatGPT. His endorsement has fueled renewed speculation that Google may now be emerging as the frontrunner in the AI race—after years of trailing OpenAI and its flagship product, ChatGPT. Pichai reflected on Google’s long-term AI strategy, recalling his vision from 2016 to make the entire company AI-first. He pointed to foundational milestones over the past decade: the creation of Google Brain in 2012, the acquisition of DeepMind in 2014, AlphaGo’s historic victory over a top human Go player, and the development of Google’s custom tensor processing units (TPUs) used to train models like Gemini. “That was a full-stack bet,” Pichai said, “to position Google as an AI-first company.” He emphasized that while the company may have appeared quiet or slow to outsiders during the early days of generative AI, it was actively building the infrastructure and expertise needed to compete at scale. The launch of Gemini represented the culmination of that effort. Google consolidated its AI talent from Google Brain and DeepMind, expanded its AI infrastructure, and accelerated development across all layers of the stack—from hardware and training to post-training optimization and real-time computation. Pichai described this full-stack approach as central to Google’s AI ambitions: improving every component of the system to deliver smarter, faster, and more efficient models. He acknowledged that building this foundation took time and investment, especially when the company initially lacked the capacity to respond quickly to the generative AI boom. “If you were on the outside, it would look like we were quiet, or behind,” he said. “But we were putting all the building blocks in place—and then executing on top of them.” Now, he said, the tide has turned. “We're on the other side now.”