Networks & InfrastructureHpeAmdMunish MehtaAI Networking
HPE hires AMD network chief Munish Mehta for AI networking push
HPE names AMD global network head Munish Mehta VP of product management for data center networking, tasking him with open networking products for AI infrastructure.
Networks & InfrastructureWhy it matters
- Munish Mehta joins HPE as VP of product management for data center networking after nearly three years at AMD.
- Mehta claims over $10 million in savings and 25-fold data center switching capacity gains at AMD.
- He spent nine years in a second Cisco stint, becoming VP of engineering in 2021 and leading Nexus 9000 and ACI work.
- Mehta says he will lead HPE open networking products for AI infrastructure.
The story
Hewlett Packard Enterprise has hired AMD's global network head Munish Mehta as vice president of product management for data center networking, a move that puts a two-decade networking engineer at the center of HPE's AI infrastructure strategy.
Mehta arrives after just under three years at AMD, where he ran worldwide WAN, campus, cloud and data center networks as global head of network and data center infrastructure services. He also oversaw deployment of AMD-silicon networking into production at the chip maker.
His record there, as stated on his LinkedIn profile, is dense with numbers. He nearly doubled the size of the networking team. He claims more than $10 million in savings through architectural consolidation and price negotiation. He also says he separated machine-to-machine traffic from user traffic to quadruple WAN capacity, cut SD-WAN critical incidents to near zero, and raised data center switching and wireless capacity 25-fold and 10-fold respectively.
What does the hire change for HPE?
The appointment signals where HPE sees its next revenue growth: open networking products for AI infrastructure. Mehta said as much in his LinkedIn post announcing the move, describing AI networking as central to HPE's growth strategy.
HPE, formed from the 2015 split of Hewlett-Packard and now anchored by its Juniper Networks acquisition, has been assembling a full AI stack — compute, storage and networking — to compete for enterprise and hyperscaler build-outs. A product management leader who has bought, operated and engineered data center fabrics addresses a gap between HPE's portfolio breadth and its need for products operators can actually run at scale.
"I've spent my career on both sides of the data center network," Mehta wrote. "At Cisco, I led the engineering teams behind Nexus 9000, ACI, and Nexus Dashboard. At AMD, I became the customer, running global networks and data centers through the company's AI build-out. Both taught me what AI infrastructure teams need: open, high-performance fabrics that are simple to operate at scale."
He added: "HPE now brings together the complete AI infrastructure portfolio: compute, storage, and networking. The leadership team is focused, hungry for growth, and committed to execution."
A career built on SDN and data center fabrics
Mehta's engineering pedigree runs deep. He joined Cisco in 1999, left for Big Switch Networks in 2012, then moved to Insieme Networks, the Cisco spin-in acquired by Cisco in 2014. His second Cisco stint lasted nine years and ended with him as vice president of engineering in 2021. There he led teams delivering Cisco's SDN ACI controller for automation of hybrid cloud networking infrastructure, along with the Nexus 9000 switch line and Nexus Dashboard named in his post.
At AMD he also conceived and led SIGMA, an agentic AI platform he says was adopted by hundreds of IT engineers. The platform, per his account, provides unified visibility, analytics and automation across heterogeneous multivendor networks, telecom spend, data center space and power, and enterprise applications. SDxCentral found no public documentation of the platform in a preliminary search, so its capabilities rest on Mehta's own description.
What comes next
No start date beyond the LinkedIn announcement was disclosed, and HPE has not detailed which product lines Mehta will own. Given his stated mandate — open networking products for AI infrastructure — his first public roadmap decisions are likely to land in HPE's data center fabric portfolio, where the company is betting that simpler, open fabrics will win AI cluster deals against incumbent closed ecosystems.
Also reported
Source: sdxcentral.com
More from James Calloway
Show full bio
Staff writer covering companies, deals and strategy in Carrier business, 5G and 6G, fiber and broadband at Telecom Gazette.
499 articles