MPN: Z9664F-ON·By: Dell
Ships from the US.
We'll confirm availability and pricing for Z9664F-ON, usually within one business day.
We'll confirm availability and pricing for Z9664F-ON, usually within one business day.
The Dell PowerSwitch Z9664F-ON is a high-density 2RU aggregation switch designed for modern data centers requiring extreme 400GbE performance. It delivers 51.2Tbps full-duplex switching capacity and 10.2Bpps forwarding rate across 64 QSFP56-DD ports, supporting native 400GbE or breakout to 256x100GbE, 10/25/40/50/200G speeds. With sub-850ns latency and 113.66MB packet buffer, it ensures non-blocking wire-speed performance for AI/ML, HPC, and cloud workloads sensitive to packet loss. Key features include open-source ONIE for zero-touch OS installation (SmartFabric OS10, SONiC), hot-swappable redundant power supplies (up to 2125W max), 4+1 fans, and flexible airflow (I/O-to-PSU or PSU-to-I/O). Multi-rate optics/cables compatibility simplifies migrations, while robust L2/L3 features like 156K MAC table, 4K VLANs, and PTP/1588 timing enable scalable spine/leaf architectures. Ideal for reducing network layers in distributed AI training or high-throughput storage fabrics.
| Manufacturer part # | Z9664F-ON |
|---|---|
| Brand | Dell |
| Condition | New |
| Model | Z9664F-ON |
| VLANs | 4K (MSTP: 64 instances) |
| Timing | PTP/1588, Sync-E compliant |
| Latency | <850ns port-to-port |
| NOS Support | ONIE: OS10, SONiC, others |
| Product Line | PowerSwitch Z-Series |
| Product Type | 400GbE Aggregation Switch |
| Packet Buffer | 113.66MB (56.83MB shared) |
| Certifications |
Native ONIE support for major NOS; broad multi-rate optics/cable compatibility with Dell AI fabrics.
Direct 400GbE to GPU clusters
Provides non-blocking 51.2Tbps backbone connectivity for leaf-spine fabrics handling east-west AI traffic.
Direct connects up to 256 GPU nodes at 400GbE/100GbE, minimizing latency in distributed fine-tuning clusters.
Sub-850ns latency and 10.2Bpps forwarding support high-performance computing workloads with PTP timing.
Multi-rate ports with ONIE/SONiC enable scalable, disaggregated networking for hyperscale deployments.
Breakout to 256x100GbE optimizes NVMe-oF/SAN fabrics with low packet loss tolerance.
Compact 2RU design with redundant components fits dense ToR roles in hybrid cloud setups.
Details
Shipping, warranty & returns
No reviews yet — be the first to review this part.
Secure SSL checkout. Visa, Mastercard, Amex and Discover accepted. ACH / wire available for businesses.
New, certified-refreshed and OEM stock with guaranteed compatibility, backed by warranty on every part.
We process quotes, POs, stock documentation and compliance invoicing.
[email protected]| RoHS, TAA options |
| Optics Support | 400G QSFP56-DD SR8/FR4/LR4/ER4/ZR; legacy QSFP28/SFP |
|---|
| MAC Address Table | 156K entries |
|---|
| Max Ports via Breakout | 256x 100GbE |
|---|
| Airflow | I/O Panel to PSU or PSU to I/O |
|---|---|
| Forwarding Rate | 10.2Bpps |
| Switching Capacity | 51.2Tbps full-duplex non-blocking |
| Max Power | 2125W; Typical 500W |
|---|---|
| Power Supplies | 2x hot-swap redundant AC 100-240V 50/60Hz |
| Fans | 4x hot-swap (3+1 redundancy) |
|---|---|
| Weight | 23kg (50.71lbs) with PSUs/fans |
| Form Factor | 2RU Rackmount (17.26x24.80x3.45in / 438.5x630x87.5mm) |
| Operating Temp | 0-45°C (32-113°F) |
|---|
| Ports | 64x 400GbE QSFP56-DD multi-rate (10/25/40/50/100/200/400G) |
|---|
The Z9664F-ON offers 51.2Tbps full-duplex switching fabric and 10.2Bpps forwarding across 64x 400GbE QSFP56-DD ports. Ports are multi-rate, supporting 10/25/40/50/100/200/400G via breakout cables for up to 256x100GbE, ideal for dense data center aggregation with non-blocking performance.
Yes, it supports ONIE for zero-touch provisioning of Dell SmartFabric OS10, SONiC, or third-party NOS. This disaggregated architecture allows flexible OS selection, simplifying migrations to software-defined networks in cloud or AI environments.
Max power draw is 2125W (typ. 500W) with 100-240VAC PSUs; operates 32-113°F (0-45°C). Features 2RU chassis (17.26x24.80x3.45in), redundant 3+1 hot-swap fans, and dual airflow options (I/O-to-PSU or reverse) for hot/cold aisle data centers.
With sub-850ns latency, 113.66MB buffer, and direct 400GbE to 256 nodes, it excels in low-latency AI distributed training. PTP/1588 sync and non-blocking fabric reduce layers, optimizing GPU interconnects for fine-tuning large models.
QSFP56-DD ports support 400GbE SR8/SR4.2/FR4/LR4/ER4/ZR transceivers, plus legacy 200/100/40GbE QSFP. Breakout DAC/AOC for 4x100G/2x200G; SFP56-DD for 100/25/10GbE. Full compatibility list in Dell spec sheet.
Hot-swappable redundant AC/DC PSUs (2x) and 4-fan modules (3+1). Delivers non-blocking L2/L3 forwarding (156K MAC, 4K VLANs) with packet loss protection, trusted platform module, and RoHS compliance for mission-critical deployments.
You might also want to check out these products.

Cisco Meraki MS125-48LP-HW

Cisco Meraki MS125-48LP

AT-8624T/2M - Allied Telesis 8600 24-Port 2x Module Bays uplink Switch

AT-9000/24 - Allied Telesis 9000/52-10 24-Port Switch

469-3414 - Dell PowerConnect 5524 24-Port + 2 x SFP+ Ports 1U Switch

AT-GS910/24-50 - Allied Telesis CentreCOM GS910 24-Ports Switch

AT-FS717FC/SC-10 - Allied Telesis 16x 1x 100BFX SC Connector Switch

AT-9424T/SP - Allied Telesis 9400 24-Port 2x Port SFP Switch