Google semiconductor and AI chip development

Google is advancing its custom semiconductor development through new hardware launches and regional expansion.

In August 2026, Google and Intel introduced the E2000, a co-designed infrastructure processing unit (IPU) designed to manage data-intensive cloud workloads more efficiently. This chip is integrated into Google Cloud’s C3 Virtual Machines. During this period, Google also developed the Tensor G6 chip for the Pixel 11 series, which transitioned to Arm’s latest mobile cores.

By September 2026, Google expanded its semiconductor efforts in Israel by forming a specialized team focused on artificial intelligence chips. This initiative, led by former Hailo executive Guy Kaminitz, aims to develop hardware capable of running AI tasks directly on devices to reduce dependency on cloud servers. The expansion includes recruiting talent from companies such as NVIDIA and Pliops to support Google’s growing engineering presence in Tel Aviv and Haifa.

Reports from Haaretz indicate this new Israeli team is specifically targeting the development of AI chips for robots and autonomous systems. This move seeks to accelerate processing speeds by enabling on-device task execution. Google’s chip development operations in Israel, which began through its Google Cloud division in 2021, now include approximately 700 engineers across Tel Aviv and Haifa working on computing infrastructure and chip design. Google has declined to comment on these specific developments.