On the {hardware} entrance, HPE is focusing on the AI knowledge heart edge with a brand new MX router and the scale-out networking supply with a brand new QFX change. Juniper’s MX sequence is its flagship routing household geared toward carriers, large-scale enterprise knowledge heart and WAN prospects, whereas the QFX line companies knowledge heart prospects anchoring backbone/leaf networks in addition to top-of-rack methods.
The brand new 1U, 1.6Tbps MX301 multiservice edge router, obtainable now, is geared toward bringing AI inferencing nearer to the supply of knowledge era and might be positioned in metro, cell backhaul, and enterprise routing functions, Rahim stated. It consists of high-density help for 16 x 1/1025/50GbE, 10 x 100Gb and four x 400Gb interfaces.
“The MX301 is basically the on-ramp to offer excessive pace, safe connections from distributed inference cluster customers, units and brokers from the sting all the way in which to the AI knowledge heart,” Rami stated. “The necessities listed here are sometimes round excessive efficiency, but additionally very excessive logical abilities and built-in safety.”
Within the QFX enviornment, the brand new QFX5250 change, obtainable in 1Q 2026, is a completely liquid-cooled field geared toward tying collectively Nvidia Rubin and/or AMD MI400 GPUs for AI consumption throughout the info heart. It’s constructed on Broadcom Tomahawk 6 silicon and helps as much as 102.4Tbps Ethernet bandwidth, Rahim stated.
“The QFX5250 combines HPE liquid cooling know-how with Juniper networking software program (Junos) and built-in AIops intelligence to ship a high-performance, power-efficient and simplified operations for next-generation AI inference,” Rami stated.
Partnership expansions
Additionally key to HPE/Juniper’s AI networking plans are its partnerships with Nvidia and AMD. The corporate introduced its relationship with Nvidia now consists of HPE Juniper edge onramp and long-haul knowledge heart interconnect (DCI) help in its Nvidia AI Computing by HPE portfolio. This extension makes use of the MX and Junipers PTX hyperscaler routers to help high-scale, safe and low-latency connections from customers, units, and brokers to AI factories and connections between clusters deployed throughout longer distances or throughout a number of clouds, Rahim stated.
Leave a Reply