“Introducing Axion: Google’s Revolutionary Arm-based Data Center Processor”

Google Cloud has finally joined its competitors, Amazon Web Services (AWS) and Microsoft Azure, in unveiling its very first custom-built Arm processor called Axion.

This new processor is based on Arm’s Neoverse 2 designs and, according to Google, it offers 30% better performance compared to other Arm-based instances from AWS and Microsoft. It also boasts 50% better performance and 60% better energy efficiency compared to traditional X86-based instances.

However, Google has not provided any documentation to back up these bold claims. When questioned for more information, the company politely declined to provide any additional details. Availability dates, pricing, and technical data were all left unanswered. Even the supposed “benchmark” results were not supported by the name of the X86 instance used for comparison.

According to Google spokesperson, Amanda Lam, “Technical documentation, including benchmarking and architecture details, will be available later this year.”

One might wonder if these chips are even ready for use yet. After all, it took Google quite a while to announce Arm chips in the cloud, especially when considering their track record of developing in-house TPU AI chips and custom Arm-based mobile chips for their Pixel phones. In contrast, AWS launched its Graviton chips as early as 2018.

However, it’s worth noting that Microsoft announced its Cobalt Arm chips only late last year and they are still not available to customers. Meanwhile, Microsoft Azure has been offering instances based on Ampere’s Arm servers since 2022.

During a press briefing prior to the official announcement on Tuesday, Google emphasized that since Axion is built on an open foundation, it allows their customers to seamlessly bring their existing Arm workloads to Google Cloud without any modifications. This is to be expected, as anything else would have been a major setback for Google Cloud.

Mark Lohmeyer, Google Cloud’s Vice President for compute and AI/ML infrastructure, further explained, “We recently contributed to the SystemReady Virtual Environment, which is Arm’s hardware and firmware interoperability standard that ensures common operating systems and software packages can run seamlessly in ARM-based systems. Through this collaboration, we have access to a broad ecosystem of cloud customers who have already deployed ARM-based workloads across hundreds of ISVs and open-source projects.”

More information on Axion is expected to be revealed later this year.