Skip to content
Product Overview

Transceivers for every tier, from 1.6T to 100G.

MSA-compliant optics spanning the highest-bandwidth AI/ML fabrics down to the data center workhorse tiers — OEM-agnostic across every major platform, coded and tested on live switching hardware.

5,000+SKUs
8MSA Standards
300+OEM Environments Replicated

Cisco, Arista, Juniper, Mellanox, HPE, Dell, Huawei, Nokia, and more — OEM-agnostic across every major platform.

Coded to your platform
In-system tested
Lifetime warranty
Same-day shipping
1.6T Transceiver
1.6T

1.6T Transceivers

1.6T transceivers are the newest generation of optical interconnect, built for hyperscale AI and machine learning clusters that demand extreme bandwidth density between GPU fabrics and switch spines. They target next-generation 802.3dj-based Ethernet architectures and are used by hyperscalers and AI infrastructure providers scaling GPU-to-GPU and GPU-to-switch fabric bandwidth beyond what 800G can support. Available in emerging OSFP-XD and similar next-gen form factors as the ecosystem matures.

Angled view of an 800G OSFP transceiver with a large yellow pull tab, a silver metal housing with a finned end and gold edge connector, and a white label with a barcode, QR code and small circular logo.
800G

800G Transceivers

800G transceivers deliver the bandwidth density required for AI/ML training clusters, hyperscale spine layers, and next-generation data center fabrics. Built on OSFP and QSFP-DD800 form factors using PAM4 encoding, they double the throughput of 400G optics while maintaining rack density. Data center architects use 800G to reduce the number of physical links needed in high-bandwidth AI fabric and core network designs.

Angled view of a 400G QSFP-DD transceiver with a blue pull tab marked 400G, a silver housing with a finned section near the tab end, a gold edge connector at the far end, and a white label with a barcode, QR code and small circular logo.
400G

400G Transceivers

400G transceivers (QSFP-DD and OSFP) are the current standard for hyperscale and large enterprise data center backbones, defined under IEEE 802.3bs. They use PAM4 encoding to quadruple 100G throughput per port and are deployed in leaf-spine core layers, service provider backbones, and high-density interconnects. Network architects choose 400G when scaling core bandwidth without proportionally increasing the number of physical ports.

Angled view of a 100G QSFP28 transceiver with a blue pull tab, a silver metal housing, a gold edge connector at the far end, and a white label with a barcode, QR code and small circular logo.
100G

100G Transceivers

100G QSFP28 and SFP56 transceivers are the workhorse of modern data center leaf-spine architectures, service provider aggregation layers, and 100G Ethernet/OTU4 backbones. QSFP28 covers the standard four-lane deployments, while single-lane SFP56 fits higher-density, lower-power server and switch ports. They support reaches from under 100 meters to 80 kilometers depending on variant, over both multi-mode and single-mode fiber. IT teams standardize on 100G for server-to-spine and inter-switch links where 400G density isn't yet needed but 40G no longer has enough headroom.

Angled view of a 40G QSFP transceiver with a yellow pull tab, a silver metal housing, a gold edge connector at the far end, and a white label with a barcode, QR code and small circular logo.
40G

40G Transceivers

40G QSFP+ transceivers bridge the gap between legacy 10G infrastructure and modern 100G+ backbones, commonly deployed in mid-size data centers, metro area networks, and inter-switch links needing more headroom than 10G but not yet requiring 100G. Compliant with IEEE 802.3bm, they support reach options from under 150 meters to 40 kilometers over LC or MPO interfaces. Many enterprises still run 40G in core aggregation layers where a full 100G upgrade isn't yet cost-justified.

25G Transceiver
25G

25G SFP28 Transceivers

25G SFP28 transceivers are the standard choice for modern top-of-rack server connectivity, offering more than double the bandwidth of 10G in the same physical form factor. Evolved from 100 Gigabit Ethernet's four-lane architecture, each 25G link carries a quarter of a 100G connection's bandwidth over a single lane. Data center operators use 25G as the cost-efficient upgrade path from 10G server links without needing new switch hardware footprints.

10G Transceiver
10G

10G Transceivers (SFP+)

10G SFP+ transceivers remain the most widely deployed interconnect in enterprise networks, powering access-layer switching, Fibre Channel SAN connectivity, and metro/telecom links. Compliant with 10 Gigabit Ethernet, 10G Fibre Channel, and OTN standards, they support digital diagnostics for real-time link health monitoring. IT teams rely on 10G optics as the dependable, cost-effective standard for enterprise networks that don't yet require 25G or higher.

Angled view of a 1G SFP transceiver with an RJ45 port and gold-colored latch at one end, a gold edge connector at the other, and a white label with a barcode, QR code and small circular logo.
1G

≤1G Transceivers (SFP)

1G SFP transceivers remain essential for legacy switch access ports, out-of-band management networks, PON deployments, and edge connectivity in enterprise and telecom environments. As one of the longest-standing pluggable optical standards, SFP supports Gigabit Ethernet, Fibre Channel, and SONET across a wide range of reach options. Organizations maintaining older switching infrastructure or building out low-bandwidth edge and management networks depend on 1G SFP for reliable, hot-swappable connectivity.

Cabling and accessories

DAC / AOC cables
Cabling

DAC and AOC cables

A DAC (direct attach copper) cable and an AOC (active optical cable) are both pre-terminated cables with a pluggable connector on each end. A DAC carries the signal over twinaxial copper and suits short runs, usually inside a rack or to the next one. An AOC carries it over fiber with optics built into each end and suits longer runs and tighter cable pathways.

Fiber patch cables
Cabling

Fiber patch cables

A fiber patch cable (or patch cord) is a length of optical fiber with a connector on each end, used to link a transceiver to a patch panel or to another device. To pick the right one, match four things: the fiber type (single-mode or multimode), the connector (LC, SC or MPO), the polish, and the number of fibers.

Copper cabling
Cabling

Copper cabling

Copper network cabling uses twisted pairs and RJ45 connectors for Ethernet links from Gigabit to 40 Gigabit. The category number sets the maximum speed and distance: Cat5e for 1G, Cat6 and Cat6a for 10G, and Cat8 for 25G and 40G over short runs. Pick the category for the fastest link you expect to run over that cable, not just today’s speed.

Cable management
Cabling

Cable management

Cable management is the hardware and practice that keeps dense fiber and copper runs organized, protected and serviceable. In a rack that means patch panels to terminate runs, horizontal and vertical managers to route patch cords, and ties or straps to bundle them without crushing the cable.