Axion also leverages more than a decade of Google\u2019s custom silicon innovation. This enables Google to more readily incorporate customer feedback into chip designs and address the more general, though complex, needs of CPUs.\u00a0<\/p>\n
Matching workload type to the processor<\/p>\n
Bhumik Patel<\/a>, Director of Software Ecosystem Development at Arm, says the key to all of this is to match the workload type as closely as possible to computing capacity. CPU-powered cloud instances are a practical option for certain AI workloads, particularly those with smaller datasets or less complex models.\u00a0<\/p>\n \u201cAgentic tasks such as orchestrating, talking to APIs, and memory management are all ones CPUs are good at, so it\u2019s a distributed and concurrent AI workload,\u201d Patel tells The New Stack. Intelligent workload-processing apportionment makes agentic AI more cost-effective and efficient than running all workloads on a single compute type.<\/p>\n This efficiency is quantifiable. The Google Kubernetes Engine Agent Sandbox<\/a> running on Google Axion N4A provides up to 30% better price performance<\/a> than the next hyperscale cloud provider, says Google\u2019s Mo Farhat<\/a>, Axion Group Product Manager. The GKE Sandbox is an open-source Kubernetes-native primitive designed to execute untrusted AI-generated code safely.\u00a0<\/p>\n \u201cAgentic tasks such as orchestrating, talking to APIs, and memory management are all ones CPUs are good at, so it\u2019s a distributed and concurrent AI workload.\u201d<\/p>\n Intelligent workload decoupling makes agentic AI significantly more cost-effective. Google Cloud\u2019s fluid computing foundation enables engineering teams to reserve specialized accelerators strictly for heavy reasoning and generative workloads, while leveraging Axion CPUs for high-concurrency orchestration and context management.<\/p>\n Secure execution with the GKE Agent Sandbox<\/p>\n As agents begin to generate and execute dynamic code autonomously, security is non-negotiable. Running AI-generated code directly in a standard cluster poses severe security risks, as untrusted code could potentially access other apps or the underlying cluster node.<\/p>\n The Google Kubernetes Engine (GKE) Agent Sandbox resolves this by providing an isolated environment for safely executing untrusted code. Running on Axion-powered N4A instances, the sandbox provides up to 30% better price performance than comparable workloads on other hyperscalers.<\/p>\n The vertical stack isolates sensitive tasks at the kernel level with sub-second latency.<\/p>\n The vertical stack isolates sensitive tasks at the kernel level with sub-second latency.\u00a0 GKE Agent Sandbox natively supports gVisor<\/a> (an open-source application kernel developed by Google that acts as a secure sandbox for containers) and default-deny Kubernetes network policy. Agent Sandbox provides pluggable interfaces for open-source sandboxes, such as Kata Containers<\/a>, enabling users to customize their kernel isolation.\u00a0<\/p>\n Powered by gVisor technologies with software support from Arm\u2019s architecture, the sandboxes intercept and validate system calls before they reach the host kernel. These isolated execution environments enable deployment of autonomous systems at scale without sacrificing performance or operational agility.<\/p>\n To manage resources efficiently when agents sit idle, GKE Pod snapshots allow users to save and restore the exact process state of sandboxed environments. This functionality provides four major architectural benefits:<\/p>\n Fast startup: Reduces sandbox startup time by restoring from a pre-warmed snapshot rather than initializing from scratch.<\/p>\n Long-running agents: Pauses sandboxes that take a long time to run and resumes them later\u2014or moves them across nodes\u2014without losing progress.<\/p>\n Stateful workloads: Persist an agent\u2019s context, such as conversation history or intermediate calculations.<\/p>\n Reproducibility: Captures a specific state to use as a baseline for spinning up multiple new sandboxes.<\/p>\n Getting started<\/p>\n As token generation, autonomous workflows, and continuous agent interactions grow exponentially, relying exclusively on accelerator-backed stacks for every task will become financially and architecturally unsustainable.<\/p>\n The combination of CPU and accelerator execution accounts for bursts in agent activity and unpredictable demand spikes by eliminating the inference tax. Google Cloud\u2019s full-stack advantage enables organizations to deploy the right machine for the job.\u00a0<\/p>\n By using Google Axion and GKE Agent Sandbox, builders can optimize total cost of ownership and security while maintaining the performance required for AI agents.<\/p>\n Learn more about Google Axion<\/a>.<\/p>\n
![](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/07\/9056536e-cropped-a44aa3a7-1597405674814-600x600.jpeg\"\/)
<\/p>\n