Guides
Server and rack integration, explained.
What L10 and L11 actually include, how each level is tested, and how to choose between them. Written for the people specifying and buying AI infrastructure.
Guide · Rack integration
What is L11 rack integration?
L11 rack integration explained: what's racked, cabled and tested before an AI rack ships, the handover documents to ask for, power and cooling limits, and how L11 differs from L10 and L12.
Read the guideGuide · Server integration
What is L10 server integration?
L10 server integration explained: how a barebone becomes a configured, firmware-matched, burn-in-tested GPU server, what's tested, what L10 doesn't include, and when it's all you need.
Read the guideGuide · Comparison
L10 vs L11 integration
L10 ships a tested server; L11 ships a tested rack. A side-by-side of scope, testing, handover, timeline and cost drivers, the full L1–L12 scale, and a checklist for choosing between them.
Read the guideThe levels TORmem builds to
| Level | What's included |
|---|---|
| L6 · Barebone | Chassis, motherboard, backplane, power supplies, fans. No CPU, memory, GPU or drives. |
| L10 · Configured node | Everything in L6, plus CPU, DDR5, GPUs, NVMe, network adapters — assembled, firmware-matched and burn-in tested. |
| L11 · Rack-integrated | Everything in L10, plus rack, PDUs, switching, optics, cabling, cluster-level validation and on-site delivery. |
Every build is specified with the client. Send a procurement brief or see L11 rack integration.