Cisco UCS M8 Series: 8th-Generation Compute Unleashed

Cisco's UCS M8 series

Cisco's UCS M8 series marks a major evolution in unified compute, combining high-performance Intel and AMD platforms across modular and rackmount designs. Built for hybrid workloads and cloud-managed through Cisco Intersight, these systems deliver powerful compute density, flexible I/O and streamlined lifecycle management.

As illustrated in the graphic below, Cisco's UCS family spans dense GPU servers, modular X-Series blades and versatile rackmount systems. 

Cisco's UCS X-Series marks a transformational shift in data center design by decoupling compute, I/O and storage into modular, hot-swappable components within a unified chassis. At the center of this innovation is the Cisco UCS X9508 chassis, a 7RU enclosure that supports up to eight horizontal compute nodes, offering flexible resource allocation and high-performance density with centralized power and cooling.

Unlike legacy blade systems that are limited by form-factor rigidity, the X9508 introduces a future-ready fabric architecture. It eliminates the midplane, allowing direct node-to-fabric interconnects and opening the door for future bandwidth upgrades and peripheral expansion without a chassis redesign. The I/O is modularized through Cisco VICs and the X-Fabric module, which allow for GPU expansion, PCIe add-ons or storage acceleration — all decoupled from the compute itself.

The roadmap for X-Series is ambitious and designed to evolve in place. Cisco has already introduced high-bandwidth, PCIe Gen5-capable I/O fabrics and is expected to integrate CXL (Compute Express Link) in upcoming revisions. This will allow memory pooling and low-latency GPU-to-CPU communication across nodes, enabling true composable infrastructure. Future nodes may support disaggregated memory, modular AI accelerators and direct-to-GPU fabrics, all while remaining compatible with the current chassis.

This vision of composability is not theoretical — Cisco is actively executing on it. With the release of both Intel and AMD-powered nodes, the X-Series becomes the most flexible compute platform Cisco has ever produced.

X210c M8: Intel Xeon performance in blade format

The X210c M8 is Cisco's latest Intel-based compute node, optimized for general-purpose enterprise workloads in high-density scenarios. It supports dual 6th-generation Intel Xeon Scalable processors, each with up to 86 cores, and provides 32 DDR5 DIMM slots, enabling memory capacity up to 8 TB per node.

Storage flexibility is enhanced with support for up to nine (9) EDSFF E3.S NVMe drives, delivering high-speed, front-accessible storage without consuming excessive space or airflow. Connectivity is driven through the VIC 15422 mLOM, enabling dual 100 GbE fabric paths per node. Additionally, a mezzanine slot allows for optional hardware expansion — typically used for GPUs or specialized NICs.

Use-case perspective: The X210c M8 is well-suited for scalable VDI, high-density virtual workloads, and multi-tenant Kubernetes clusters. When deployed in WWT's Advanced Technology Center (ATC) lab, it serves as a high-throughput environment for proofs of concept (POCs) where node-based segmentation and dynamic provisioning via Intersight are key to testing real-world outcomes.

See Cisco's spec sheet here.

X215c M8: AMD EPYC efficiency and scale

The X215c M8 brings AMD's EPYC power into the X-Series modular family. Supporting a single 4th or 5th-generation EPYC processor with up to 160 cores, this node offers excellent compute density with a smaller power and thermal footprint — an ideal complement to Intel-powered counterparts.

It includes 12 DDR5 DIMM slots, allowing up to 1.5 TB of RAM, and features the same nine E3.S NVMe bays and VIC 15422 connectivity as its Intel sibling. It leverages EPYC's integrated PCIe Gen5 lanes for internal flexibility and future expansion through the X-Fabric pathway.

Use-case perspective: The X215c M8 is ideal for cost-efficient performance scenarios, such as high-throughput container services, storage controller nodes and microservices environments. In the ATC, it provides an AMD-native option for customer validations that compare core-per-watt efficiencies or wish to test hybrid architectures.

See Cisco's spec sheet here.

Rackmount versatility: The C-Series UCS M8 portfolio

While modular platforms offer composability and scale, rackmount servers remain the workhorses of the enterprise. With the UCS M8 generation, Cisco's C-Series has evolved to deliver exceptional performance across a range of workloads, from general-purpose compute to AI-heavy analytics — all while simplifying operations through Cisco Intersight. The M8 rack models blend power, density and storage flexibility, and are available with both Intel Xeon and AMD EPYC architectures, giving customers unmatched choice for their datacenter designs.

The three rack-mount models — C220 M8, C240 M8 (Intel), and C225 M8, C245 M8 (AMD) — offer consistent operational benefits: PCIe Gen5 support, DDR5 memory, advanced Tri-Mode storage controllers, and compatibility with 10–200 GbE networking through Cisco's OCP 3.0 and VIC adapters. These servers are built for seamless integration into UCS Manager or Intersight-managed environments, enabling full visibility and lifecycle control.

Cisco UCS C220 M8: High-density 1U Intel server

The C220 M8 is a 1U dual-socket rack server powered by Intel's 6th-generation Xeon Scalable CPUs. Each socket supports up to 86 cores and 8 memory channels, totaling 4 TB of DDR5 memory across 16 DIMM slots. Despite its compact footprint, this system delivers outstanding performance per rack unit.

Storage options include a mix of NVMe, SAS and SATA with front-drive configurations supporting up to ten E3.S NVMe drives or ten 2.5" SFF drives, depending on the chassis SKU. Internally, two PCIe Gen4 risers support accelerators or networking cards, while OCP 3.0 and Cisco VICs provide flexible networking.

Use-case examples: This server excels in dense compute environments such as virtualization clusters, edge aggregation or application servers. In the ATC, the C220 M8 is used in performance validation labs for high-density deployment scenarios where space and energy efficiency are critical. With Cisco Intersight, customers can manage thousands of these servers remotely with firmware compliance, BIOS tuning and infrastructure orchestration.

See Cisco spec sheet here. 

Cisco UCS C240 M8: Scalable 2U Intel powerhouse

The C240 M8 builds upon the C220 by expanding drive and PCIe capacity within a 2U form factor. It retains support for two 6th-gen Xeon CPUs and up to 4 TB of DDR5 memory (or 8 TB with select DIMMs), but adds enhanced I/O with room for up to 24 front-facing 2.5" SFF drives or 12 LFF drives.

Internally, customers can add up to six PCIe Gen4/Gen5 cards, including GPUs, NVMe storage controllers, or additional networking interfaces. It's an ideal candidate for compute-heavy workloads that require large data sets to be processed or cached locally.

Use-case examples: Designed for scale-out storage, database workloads, large VM farms, or ML inference tasks. ATC often uses the C240 M8 in hybrid-stack test beds, simulating enterprise-scale data pipelines. With Intersight integration, performance data, thermal profiles, and firmware versions can be monitored and tuned with centralized precision — vital for maintaining consistency across distributed clusters.

See Cisco spec sheet SFF here.
See Cisco spec sheet LFF here.

Cisco UCS C225 M8: AMD EPYC Efficiency in 1U

The C225 M8 introduces AMD's high-core-count EPYC platform into a 1U rack server. With a single 4th or 5th Gen EPYC processor, it supports up to 160 cores and 1.5 TB of DDR5 memory across 12 DIMM slots. EPYC's dense PCIe architecture delivers over 160 PCIe Gen5 lanes, enabling impressive I/O within the smallest form factor in the lineup.

This model supports multiple configurations of 2.5" NVMe or SAS/SATA drives and includes OCP 3.0 for flexible networking.

Use-case examples: The C225 M8 is ideal for microservices, lightweight container orchestration or performance-per-watt sensitive environments. Our ATC lab environment often deploys this model in cost-conscious environments or for scale-out edge infrastructure, where customers want high throughput from a single-socket server without the power draw of a dual-CPU system. Combined with Intersight's profile-based deployment, it becomes a fast-launch platform for CI/CD environments or burst compute.

See Cisco's spec sheet here.

Cisco UCS C245 M8: AMD EPYC power and scale in 2U

The C245 M8 is built for enterprise-scale processing. With two AMD EPYC CPUs supporting up to 320 total cores and up to 6 TB of DDR5 memory, this server is designed for maximum throughput and concurrency. Its 2U chassis accommodates up to 24 front-facing NVMe/SAS/SATA drives, with rear bays optional, and multiple PCIe Gen5 risers for I/O expansion.

AMD's on-die PCIe architecture enables direct access to storage, GPUs and NICs with minimal latency. Enhanced platform security features like Secure Boot, SEV encryption and Platform Secure Boot are native to the C245.

Use-case examples: The C245 M8 is an ideal platform for data analytics, large databases (OLAP/OLTP), AI inference and even model training at scale. In the ATC, it is used for benchmarking CPU-intensive workloads and comparing virtualization efficiency between Intel and AMD architectures. Intersight enables full lifecycle management including thermal mapping, power budgeting and dynamic BIOS tuning — all critical in high-performance datacenter environments.

See Cisco's spec sheet here.

Unified management and lifecycle with Cisco Intersight

Whether you're deploying UCS blades in the modular X-Series or rackmount C-Series servers, Cisco Intersight serves as the centralized control plane for all lifecycle operations. From node-level provisioning and policy-based BIOS tuning to firmware governance and performance telemetry, Intersight brings consistency, automation and observability across the entire infrastructure.

Administrators can define infrastructure policies, automate deployment workflows, monitor health and thermal patterns, and enforce compliance baselines from a single cloud-based dashboard. Integration with ATC's orchestration tools accelerates proof-of-concept development and allows for rapid rollback, upgrade staging and secure deployment at scale.

A key part of UCS's management story is the use of Cisco Fabric Interconnects (FIs). These devices act as the unified northbound access point for compute infrastructure, consolidating Top-of-Rack (ToR) switching into a single pair of devices. This dramatically reduces switch port requirements, cabling complexity  and operational costs — delivering a lower total cost of ownership (TCO). For example, in a 16-node deployment, traditional ToR designs may require 64+ switch ports and cabling runs. With FIs, that can drop by half or more, thanks to integrated uplinks and unified port management.

However, Cisco Intersight Stand-Alone Mode introduces flexibility for customers that prefer to deploy rackmount C-Series servers without fabric interconnects. In this mode, each server is managed directly by Intersight over standard out-of-band management (CIMC). While lacking the unified cable and network abstraction layer of FIs, Stand-Alone Mode still delivers full policy enforcement, firmware orchestration, telemetry and API-driven provisioning — all through the cloud. It's especially valuable in edge or remote-site deployments where simplicity, cost or existing network fabrics take precedence.

See Cisco's spec sheet here.

Summary

Within WWT's ATC, Intersight plays a foundational role supporting customer test drives, workload simulation labs and full-stack validations across hybrid and multicloud architectures. Cisco's UCS M8 series delivers powerful compute density and flexibility across both blade and rack platforms. Backed by Intel and AMD innovation, integrated with Cisco Intersight, and brought to life through the ATC's full-stack services, these platforms form a modern foundation for hybrid IT — one that scales with business needs and evolves with future workloads.