Google Cloud made M4N virtual machines generally available, offering up to 6TB memory and 400 Gbps network bandwidth.

Google Cloud announced general availability of the M4N machine series in Compute Engine, its second network- and block-storage-optimized virtual-machine offering. These are CPU-based instances powered by fifth-generation Intel Xeon Scalable processors and Google’s Titanium offload architecture, rather than GPU instances.

The series offers up to 6TB of memory, memory-to-core ratios up to 26.57 GB per virtual CPU (vCPU), and 400 Gbps network bandwidth. Google targets high-memory, input/output-intensive workloads that otherwise require additional compute cores to obtain sufficient memory and storage bandwidth.

With Hyperdisk Extreme, Google specifies up to 1 million input/output operations per second (IOPS) and aggregate host storage throughput of 25,000 MiB/s, also described in the source as 25 GiB/s. Google says this doubles current M4 block-storage performance; the two throughput expressions are reproduced as supplied rather than silently treated as an exact conversion.

Listed workloads include Oracle, SAP HANA, SQL Server, IBM DB2, MySQL, PostgreSQL, healthcare systems, analytics, electronic design automation, and generative-AI retrieval-augmented generation (RAG) data layers. Google claims the highest per-core IOPS and throughput among comparable hyperscaler high-memory offerings and over 20% lower total cost of ownership for Oracle databases, but these remain vendor comparisons rather than independently verified results in the supplied material.