AMD EPYC 7763
64 cores / 128 threads of Milan generation compute — 256MB L3 cache, PCIe Gen4, the used-server sweet spot for high-density virtualization, HCI, and AI-inference hosts. Tested and shipped from Hong Kong.
Fully Tested & Documented · Hong Kong Warehouse Stock · Worldwide Shipping 3–7 Days · 1-Year Warranty
Hardware Gallery
The EPYC 7763, Up Close
EPYC 7763 — 64-Core Milan
Zen 3 architecture, 64 cores / 128 threads, 2.45GHz base and 3.5GHz boost. The workhorse of used EPYC Milan fleets — ask about matched pairs for dual-socket boards.
Tray Stock — Matched Pairs
EPYC 7763 in SP3 carrier frames, pulled from decommissioned servers. We match steppings for dual-socket builds — 128 cores / 256 threads per node.
64
cores / 128 threads (Zen 3)
256MB
L3 cache
3.5GHz
max boost
PCIe 4
128 lanes
Key Specifications
| Component | Specification |
|---|---|
| Processor | AMD EPYC 7763 (Milan) |
| Cores / Threads | 64 cores / 128 threads |
| Base / Boost Clock | 2.45GHz base / 3.5GHz max boost |
| Cache | 256MB L3 + 32MB L2 |
| Socket | SP3 (LGA 4094) |
| PCIe | PCIe Gen4, 128 lanes |
| Memory | 8-channel DDR4-3200 |
| TDP | 280W (configurable 225–280W) |
The 64-core Milan flagship. Step up to EPYC 9654 (Genoa) for DDR5 and PCIe Gen5, or down to EPYC 7543 for a lower-cost 32-core Milan part — both rotate through our AMD EPYC stock.
Transparent Pricing
2026 Reference Pricing by Configuration
We quote EPYC 7763 against live refurbished stock — Milan parts rotate fast, so we do not publish a stale index snapshot.
Single CPU
On Request
One tested EPYC 7763, matched stepping available. Tell us the board you are fitting (SP3 1P/2P) and we confirm compatibility.
Matched Pair
On Request
Two matched EPYC 7763 for dual-socket builds — same stepping for reliable NUMA behavior.
Bulk (8+ units)
On Request
Fleet refresh quantities, pulled from tested decommissioned servers. Volume pricing available.
Cross-shop with our Used Server Price Index for the Milan generation market context.
Origin
Hong Kong Warehouse
Free port — full export documentation
Transit
3–7 Days Air Freight
Tracked, insured, worldwide
Warranty
1-Year Warranty
Diagnostic report in every box
Quality Process
What We Test Before an EPYC 7763 Ships
Stepping & Date Code
Recorded per unit so matched pairs are truly matched — mixed steppings cause subtle NUMA instability.
Pin & Substrate Inspection
Every LGA4094 pad inspected under magnification for bent pins or thermal-paste contamination.
Boot & Core Test
Full 64-core POST and stress pass on a reference SP3 board — all cores, all memory channels exercised.
Thermal Behavior
Under-load temperature logged to flag units with degraded heat-spreader contact.
Firmware & Microcode
Compatible AGESA/microcode confirmed against your target board family where known.
Report in the Box
Every unit ships with its per-machine diagnostic report and warranty. Full process: inspection guide.
Frequently Asked Questions
How much does a refurbished AMD EPYC 7763 cost in 2026?
We quote against live stock rather than publishing stale numbers — Milan 64-core parts are in active rotation. Contact us for a firm quote; matched pairs are usually the best value.
What is the difference between EPYC 7763 and EPYC 7742?
Both are 64-core, but the 7763 is the newer Milan generation (Zen 3) with ~19% better IPC, 3.5GHz vs 3.4GHz boost, and PCIe Gen4. The 7742 is Rome (Zen 2) and cheaper. Choose 7763 for per-core performance, 7742 for budget.
Should I buy EPYC 7763 or EPYC 9654 (Genoa)?
The 9654 adds DDR5, PCIe Gen5, and 96 cores, but costs more and requires a newer SP5 board. The 7763 on SP3 is the mature, high-value choice for DDR4 fleets.
What workloads suit a 64-core EPYC 7763?
High-density virtualization, hyper-converged infrastructure, Kubernetes nodes, and AI-inference hosts. 64 cores / 128 threads and 256MB L3 make it a strong per-socket value.
Do you sell matched pairs for dual-socket servers?
Yes. We match steppings and verify both CPUs together on a 2P board before shipping, so your dual-socket build is stable out of the box.
Get a Current EPYC 7763 Quote
Tell us your target — virtualization, HCI, AI inference — and whether you need singles or matched pairs. We reply with availability and pricing, usually within one business day.
Request a Quote