Lenovo ThinkSystem
Lenovo ThinkSystem Rack & Tower Servers
ThinkSystem is where buyers land when the Dell and HPE quotes come back higher than the budget and the workload does not care which badge is on the bezel. The V3 generation covers the same ground as its competitors — SR630 V3 at 1U and SR650 V3 at 2U on Intel Xeon, SR655 V3 and SR665 V3 on AMD EPYC, the four-socket SR850 V3 at 2U and SR860 V3 at 4U for consolidation and in-memory databases, and the ST650 V3 tower for branch and SMB sites — and XClarity Controller gives the same class of out-of-band management the others do.
The practical caution with ThinkSystem is that the naming is terse and the configurations vary more than the model number suggests: an SR650 V3 built as a 24-bay all-flash virtualization host and one built with L40S cards for visual computing are quite different machines sharing a name. We quote the exact configuration rather than the model, which is also why a Lenovo bid is worth putting alongside the Dell and HPE ones on any refresh you are running.
This page is the general-purpose rack and tower line. Lenovo's accelerator-dense platforms — the SR675 V3 and SR670 V2, including the Neptune liquid-cooled variants — are eight-way and four-way GPU systems that need their power, cooling and fabric designed with them, so we scope those to the workload rather than listing them beside a virtualization host.
14 configurations below · quotes returned within 48 business hours · purchase orders accepted
Lenovo ThinkSystem Rack & Tower Servers we configure and supply
Every system below is quoted to your workload — accelerator count, CPU, memory, storage and fabric are specified together rather than sold as a fixed SKU.
Lenovo ThinkSystem SR650
A reliability-focused 2U server with strong price-performance for enterprise workloads.
- Competitive value
- Reliability focus
- XClarity management
Lenovo ThinkSystem SR630
A dense 1U dual-socket server for high server-count environments.
- High density at value
- Efficient operations
- Cloud-native ready
ThinkSystem SR630 V3 Dual Xeon 1U NVMe Database & Virtualization Server
Dense 1U compute for transactional databases and virtualization at scale
- Maximum rack density
- Predictable low latency
- Validated and warranty-backed
ThinkSystem SR630 V3 with 3x NVIDIA L4 1U Edge Inference Server
Compact 1U AI inference at the edge with three low-power GPUs
- Edge-ready footprint
- Energy-efficient acceleration
- Configured and tested
ThinkSystem SR650 V3 with 3x NVIDIA L40S 2U AI & Visual Computing Server
Versatile 2U platform for AI inference, fine-tuning, and visual workloads
- Workload versatility
- Strong memory headroom
- Vendor-backed reliability
ThinkSystem SR650 V3 24-Bay All-Flash 2U Virtualization Server
All-flash 2U workhorse for virtualization, VDI, and software-defined storage
- All-flash performance
- Balanced consolidation
- Tested and supported
ThinkSystem SR655 V3 Single-Socket EPYC with 6x NVIDIA L4 Inference Server
Cost-efficient single-socket EPYC node for high-throughput AI inference
- Single-socket efficiency
- High inference density
- Validated and warranty-backed
ThinkSystem SR665 V3 Dual EPYC Turin with 3x NVIDIA H100 NVL HPC Server
Dense 2U GPU compute for large-model inference and HPC acceleration
- Top-tier GPU compute
- Massive core count
- Configured and supported
Lenovo ThinkSystem SR850 V3 — 4-Socket High-Memory (8TB DDR5, SAP HANA Certified)
Four-socket, 8TB-class memory density for in-memory databases and SAP HANA.
- Massive memory footprint
- Dense 2U efficiency
- Certified and tested
Lenovo ThinkSystem SR850 V3 — 4-Socket Virtualization Consolidation Host
Consolidate dozens of VMs per node with four-socket core density in 2U.
- Fewer, denser nodes
- License efficiency
- Resilient by design
Lenovo ThinkSystem SR860 V3 — 4-Socket with 8x NVIDIA H100 NVL
Eight H100 NVL accelerators and four sockets for enterprise AI model training.
- Dense training throughput
- Balanced CPU-to-GPU feed
- Enterprise serviceability
Lenovo ThinkSystem SR860 V3 — 4-Socket with 4x NVIDIA L40S
Four L40S GPUs for cost-efficient inference, generative AI serving, and VDI.
- Versatile acceleration
- Efficient cost-per-stream
- Headroom to grow
Lenovo ThinkSystem ST650 V3 — GPU Tower with 2x NVIDIA L40S
Office-quiet GPU tower bringing AI inference and rendering to the edge.
- Datacenter-free deployment
- Quiet, flexible operation
- Enterprise-grade reliability
Lenovo ThinkSystem ST650 V3 — Storage-Rich SMB & Branch Tower Server
Quiet, storage-dense tower server for small business and branch offices.
- Generous local capacity
- Office-friendly footprint
- Dependable and serviceable
What we need to quote accurately
A configuration quote takes minutes when these are known and days of back-and-forth when they are not. You do not need all of them to start — send what you have.
- The workload: model sizes, training or inference, and expected concurrency
- Node count, and whether this is a single system or a scaling cluster
- Rack power and cooling available per rack, and inlet temperature
- Existing estate — the vendor and management tooling you already run
- Network fabric: InfiniBand, Ethernet, and the speed you are standardised on
- Timeline, and whether the budget is approved or being built
How we quote
- 1
Send the requirement
An email, a bill of materials, or a rough description of the workload. All three work.
- 2
We validate the configuration
Accelerator, chassis, fabric, power and cooling checked against each other before anything is priced.
- 3
Itemised quote within 48 hours
Line-by-line pricing, current availability and lead time, with warranty terms stated.
Lenovo ThinkSystem server supplier — buyer questions
Is ThinkSystem a real alternative to PowerEdge and ProLiant, or a budget option?
It is a real alternative. The processors, memory and drives are the same components the other two ship, and XClarity Controller does the job iDRAC and iLO do. What differs is the management tooling your team already knows and the support relationship you already hold — which is a genuine cost, not a snobbery. If your operations are standardised on another vendor, the switching cost can outweigh the hardware saving; if you are vendor-neutral or already run Lenovo, it very often does not. We will tell you which case you are in.
SR630 V3 or SR650 V3?
The SR630 V3 is the 1U model, the SR650 V3 the 2U. Take the 1U when the node is a hypervisor host that needs network, boot media and rack density, and the 2U when you need drive bays, larger PCIe cards, or headroom for accelerators later. The AMD equivalents — SR655 V3 single-socket and SR665 V3 dual-socket — are worth pricing in parallel when the workload is core-hungry or the software is licensed per socket rather than per core.
When do we need the four-socket SR850 V3 or SR860 V3?
When you are buying sockets for memory capacity rather than cores — consolidation targets and large in-memory databases are the usual reasons. The SR850 V3 is the 2U four-socket model and the SR860 V3 the 4U one, with the extra height buying expansion and drive capacity. Below that memory threshold, two dual-socket SR650 V3 nodes are normally the better buy on both capital and resilience.
Can you quote ThinkSystem against an existing Dell or HPE refresh?
Yes, and it is worth doing. Send us the configuration you have already been quoted — sockets, memory, drives, controller, network — and we will build the equivalent ThinkSystem and hand you both, with the differences that matter called out rather than hidden in a spec table. If the answer is that you should stay where you are, we will say so.
Compare with other platforms
Dell PowerEdge
Dell PowerEdge Rack & Tower Servers
View systemsHPE ProLiant
HPE ProLiant Rack & Tower Servers
View systemsSupermicro
Supermicro Twin, Blade & Storage Servers
View systemsBrowse the full range of GPU servers and AI clusters, or the component catalogue for drives, memory, optics and spares.