Google will bring ‘custom’ Intel Skylake Xeon chips to Compute Engine in early 2017

Google today announced that it will bring new “Custom Cloud” Intel Skylake Xeon chips to its Google Compute Engine public cloud infrastructure service in early 2017. Those chips form the basis of the virtual machines (VMs) and containers in which third developers can run the code for their applications and pay by the minute.

Intel recently said new Skylake Xeon server chips would be coming. But Intel likes to work closely with cloud providers. Yesterday Google and Intel announced a strategic alliance around cloud, and now Google is committing to bringing this new generation of chips to customers.

“The upcoming Xeon processor is an excellent choice for graphics rendering, simulations and any CPU intensive workload,” Google Cloud Platform product manager Hanan Youssef wrote in a blog post. “At launch, Compute Engine customers will be able to utilize the processor’s AVX-512 extensions to optimize their enterprise-class and HPC workloads.”

Just as people want to buy computers and phones with newer generations of chips, cloud providers want to provide servers packing the latest chips with the best performance and power efficiency. So that part makes sense. But the chips being talked about today are described as Custom Cloud, which is a distinction Intel started making in 2014 that represents a higher degree of collaboration.

Public cloud market leader Amazon Web Services (AWS) announced the M4 VM instance type last year that relies on “custom” Xeon Haswell chips. But for years AWS has talked up its partnership with Intel, often onstage at its annual re:Invent conference. That event is coming up in a couple of weeks. The timing of today’s announcement could be motivated by a desire at Google to get out in front of another announcement about collaboration between AWS and Intel.

In any case, Google is letting customers sign up to be informed about beta testing opportunities for the Skylake-based VMs.