New taipei: Taiwan is playing a "central role" in Google's transition toward AI-based computing, as well as in the broader industry shift, Rick Osterloh, senior vice president of Devices and Services at Google, said in a recent interview with CNA. Osterloh praised the company's Taiwan team for its "amazing RandD capabilities" and highlighted key partners such as MediaTek and Taiwan Semiconductor Manufacturing Co. (TSMC).
According to Focus Taiwan, Osterloh emphasized the importance of the region for Google's global innovation strategy, describing it as central to the shift in the AI computing landscape. Google's hardware engineering center in New Taipei, the company's largest hardware RandD center outside the United States, plays a crucial role in this transition.
Elmer Peng, Google's vice president of hardware, also highlighted Taiwan's robust ecosystem, citing its deep talent pool in silicon design, devices, and Android development, as well as the company's data center in Changhua. He noted that this ecosystem aligns perfectly with Google's "full-stack" strategy, which aims to provide comprehensive solutions across all layers of a technology stack.
Osterloh elaborated on Google's full-stack approach to AI, which spans applications, AI models, and underlying technology, including data centers and chips. By unifying its ecosystem around the Gemini family, Google enables the same core technology to be utilized across various platforms, including enterprise services in Google Cloud, automotive systems in Android Auto, and consumer devices like Android smartphones.
For users, this integration allows shared context and memory across platforms, positioning Google as "the only real full-stack provider of AI" capable of supporting both enterprises and consumers. Osterloh also announced the launch of a new capability called "Personal Intelligence" in Taiwan on April 15. This feature leverages a user's interaction history across services like Gmail and Google Photos to deliver more personalized responses and actions.
Looking forward, Osterloh noted that the first generation of "AI-native" devices will need to enhance existing functions while incorporating new AI-driven capabilities. He referenced developments such as "Gemini Task Automation," which enables automated workflows like booking a ride, and the integration of Gemini into the Chrome browser to assist users in real time.
He cautioned that creating fully AI-native devices will necessitate rethinking both user interfaces and underlying technologies to transition from an application-driven to an AI-driven computing model.