Google has introduced its seventh-generation AI chip, named Ironwood, aimed at enhancing the performance of AI applications. This chip is specifically designed for inference computing, which involves rapid calculations necessary for responding to user queries in applications like ChatGPT. The Ironwood chip represents a significant advancement in Google’s multi-billion dollar effort to create viable alternatives to Nvidia’s AI processors. It can operate in configurations of up to 9,216 chips and combines features from previous designs, increasing memory capacity to better support AI applications.
The new chip reportedly offers double the performance of the previous Trillium chip while consuming less energy. This development is part of Google’s strategy to bolster its AI capabilities, particularly through its proprietary tensor processing units (TPUs), which are used internally and via its cloud services. The company has not disclosed the manufacturing partner for the Ironwood chip. Following the announcement, Alphabet’s shares experienced a notable increase, closing up 9.7%.
