GCP – C4A VMs now GA: Our first custom Arm-based Axion CPU

At Google Next ‘24, we announced Google Axion Processors, our first custom Arm®-based CPUs designed for the data center. Today, we’re thrilled to announce the general availability of C4A virtual machines, the first Axion-based VM series, with up to 10% better price-performance than the latest generation Arm-based instances available from leading cloud providers. C4A VMs are a great option for a variety of general-purpose workloads like web and app servers, containerized microservices, open-source databases, in-memory caches, data analytics engines, media processing, and AI inference applications. 

Underpinned by Titanium offload technology and advanced maintenance capabilities, C4A is designed to power your most demanding workloads with up to 65% better price-performance and up to 60% better energy-efficiency than comparable current-generation x86-based instances.¹

These efficiency and performance gains are why key Google services such as Bigtable, Spanner, BigQuery, F1 Query, Blobstore, Pub/Sub, Google Earth Engine, and the YouTube Ads platform have already started deploying Axion-based servers in production.

“Spanner is one of the most critical and complex services at Google, powering products including YouTube, Gmail, and Google Ads. In our initial tests on Axion processors, we’ve observed up to 60% better query performance per vCPU over prior generation servers. As we scale out our footprint, we expect this to translate to a more stable and responsive experience for our users, even under the most demanding conditions.” – Andi Gutmans, VP/GM Databases, Google

Google Cloud customers can use C4A in many services including Google Compute Engine, Google Kubernetes Engine (GKE), Batch, and Dataproc. C4A VMs are also now available in preview in Dataflow, with support for CloudSQL, AlloyDB, and other services coming soon. C4A instances support most popular Linux operating systems including Container-Optimized OS, RHEL, SUSE Linux Enterprise Server, Ubuntu, Rocky Linux, and more. Arm-compatible software and solutions are available on the Google Cloud Marketplace, and we recently launched support for Arm-based instances migration in the Migrate to Virtual Machines service in preview.

C4A broadens our general-purpose VM portfolio, and is offered in a range of configurations:

C4A VMs also deliver the connectivity and storage performance your enterprise demands:

• High-Bandwidth Networking: Up to 50 Gbps standard bandwidth and up to 100 Gbps with Tier_1 networking to support high-traffic applications

• Hyperdisk storage: Google Cloud’s latest generation Balanced and Extreme Hyperdisk storage for scalable, high-performance storage with up to 350k IOPS and 5 GB/s throughput

What our customers and partners are saying

“At Applause, we’re committed to delivering world-class testing and feedback solutions to our global clients, and that demands a robust, efficient testing infrastructure. We’ve been impressed by Google Cloud’s Axion CPUs and C4A VMs, as they’ve shown strong performance and value across our key testing workloads. This translates to faster testing cycles, more comprehensive feedback, and ultimately, better software for our clients. We’re excited to expand our use of C4A VMs to further optimize our testing platform, supporting our clients in consistently launching high-quality digital experiences.” – Rob Mason, CTO, Applause

“Cloudfleet is on a mission to simplify Kubernetes compute infrastructure provisioning across leading cloud platforms. During the preview, we thoroughly tested the new Axion instances on our analytics workloads and build infrastructure with up to 30% improvement in performance. We now plan to migrate the rest of our workloads to Axion instances and offer Axion as part of the Cloudfleet Kubernetes Engine for our customers. We are eager to see how this will enhance their efficiency and reduce costs.” – Alexander Murazik, Head of Data, Cloudfleet.ai

“Developing AI-based applications and experiences will become the next great enterprise brand distinction. The combination of Couchbase Capella Columnar running on Google Cloud’s Axion-based instances creates instrumental value for the future’s critical, AI-powered (or agentic) applications. Our customers regularly choose Capella for its performance, versatility and scale as the database foundation for their analytical applications. And they will undoubtedly appreciate the exponential boost in performance and value of Arm-based Axion infrastructure for these workloads, especially when they need to offset the costs of incorporating AI within them.” – Matt McDonough, SVP of Product and Partners, Couchbase

“At Dailymotion, our focus is always on delivering the highest performance and efficiency for our services. As part of this commitment, we recently migrated our entire Machine Learning-powered recommendation stack to Google Cloud. Looking ahead, we are pushing the boundaries further by transitioning part of our workloads to Arm architecture. After starting with T2A instances a year ago, we are now leveraging Google’s new C4A VMs. This migration has resulted in significant improvements, including a 20% reduction in CPU consumption, while enhancing overall performance in production, and helping to reduce our carbon footprint. These promising results are driving our ambition to extend this migration across all our microservices to Arm architecture in the coming months.” – Corentin Mantion, Senior DevOps Engineer, Dailymotion

“Honeycomb.io helps engineering teams debug their production systems quickly and efficiently. Sampling is a key mechanism for controlling observability costs. For our customers who are running applications on Google Cloud, we have validated that the new Axion CPUs and C4A VMs offer the best price-performance on Google Cloud for running our Refinery sampling proxy to forward only the most important, representative samples to Honeycomb.” – Liz Fong-Jones, Field CTO, Honeycomb

“At IBM Instana, we’re committed to helping organizations achieve optimal application performance. Our automated APM platform relies heavily on real-time data processing and analysis. We’re incredibly impressed by the gains we’re seeing with Google Cloud’s Axion processors, with up to 70% responsiveness improvements and 30% higher throughput compared to prior generation VMs. We’re looking forward to deploying Instana Observability on Axion to empower our customers to build and maintain ever higher performing applications.” – Chris Bailey, CTO, IBM Instana Observability

“loveholidays is one of the fastest websites in travel. With Axion, we no longer have to trade off performance with cost optimization. We get an unprecedented level of performance showing a 40% reduction in latency against T2D instances, all whilst saving costs on capacity.” – Dmitri Lerko, Head of Engineering, loveholidays

“Paramount+ is committed to delivering the highest quality viewing experiences across all our platforms. Google Cloud’s Axion and C4A VMs provide the performance and efficiency we need to leverage the latest advancements in video encoding technology. With C4A VMs, we’re able to utilize higher efficiency of the new Arm processors to achieve 33% faster encode times, compared to older VMs. This ensures we can encode and distribute content at scale for our viewers in a timely manner, with a focus on video quality, vibrant visuals and immersive audio regardless of their device or connection.” – Jignesh Dhruv, VP Video Engineering, Paramount Streaming

“Our tests on Axion have shown roughly 250% better performance than we’ve seen on our workloads on Google Cloud to date.” – Dave Zolotusky, Principal Engineer, Spotify

“At Verve, we build meaningful connections between customers and brands through the power of data. Our platform processes massive amounts of data in real-time to deliver personalized and relevant experiences. Axion’s potential to enhance performance and efficiency is incredibly valuable to us. By leveraging Axion’s capabilities, we can further optimize our data processing, reduce latency, and ultimately deliver even more impactful and engaging experiences for our customers and their consumers.” – Prasanna Prasad, CTO, Verve

“In the fast-paced world of programmatic advertising, every millisecond counts; at Viant, we’re dedicated to providing our clients with the most efficient Demand-Side Platform (DSP) to reach their target audiences effectively. During the testing of Axion within Viant’s bidding platform, we saw an impressive 25% increase in bid processing speed, driving efficiency, and reducing costs compared to previous-generation machines. We’re excited about Axion’s potential to deliver exceptional advertising experiences while minimizing environmental impact.” – Linh Chung, CIO, Viant

Learn more

We’ve seen incredible excitement for C4A from our customers and partners so far, and we can’t wait to see what you can build with Axion, too. C4A instances are now generally available via on-demand, Spot VMs, reservations, Committed Use Discounts (CUDs) in one- and three-year terms, and FlexCUDs. C4A VMs are available in us-central1 (Iowa), us-east4 (Virginia), us-east1 (SC), eu-west1 (Belgium), eu-west4 (Netherlands), eu-west3 (Frankfurt), and asia-southeast1 (Singapore) today, with availability in additional regions coming soon. Get started on Axion today g.co/cloud/axion.

^{1. As of September 2024, based on published listed prices. Performance based on the estimated SPECrate®2017_int_base performance benchmark scores run in production on comparable latest-generation generally-available VMs with general purpose storage types.
*As of September 2024, based on published listed prices. Performance based on Google internal performance benchmark scores run in production on comparable latest-generation generally-available Intel-based VMs with general purpose storage types.}