Google has designed its very own custom chip for deep neural networks, the technology behind its artificial intelligence based internet services.
At Google I/O, CEO Sundar Pichai spoke about company’s self-designed ASIC or application-specific integrated circuit to drive deep neural networks. Google is using AI to identify objects and faces inside photos and same AI processes ‘Google Now’ voice queries.
Google is in fact using the same technology to deliver smart home speaker with Google Assistant built-in. Google calls the chip as Tensor processing unit, based on company’s TensorFlow software engine.
“We’ve been running TPUs inside our data centers for more than a year, and have found them to deliver an order of magnitude better-optimized performance per watt for machine learning,” Google said in a blog post.
Google developing its very own chip could mean trouble for traditional chipmakers like Intel who are underpinning hopes on AI as their next growth segment.