• Home
  • About
  • Events Calendar
  • Blueprint Guidelines
  • Privacy Policy
  • Manage Email Delivery
  • NextGenInfra.io
No Result
View All Result
Converge Digest
Wednesday, June 10, 2026
  • Home
  • About
  • Events Calendar
  • Blueprint Guidelines
  • Privacy Policy
  • Manage Email Delivery
  • NextGenInfra.io
No Result
View All Result
Converge Digest
No Result
View All Result

Home » Marvell Launches 260-lane PCIe 6.0 Switch

Marvell Launches 260-lane PCIe 6.0 Switch

March 18, 2026
in All
A A

Marvell introduced the Structera S 60260, which it says is the industry’s first 260-lane PCIe 6.0 switch, aimed at AI data center scale-up infrastructure. Announced at OFC 2026 in Los Angeles, the device extends Marvell’s end-to-end PCIe portfolio following its acquisition of XConn Technologies and targets one of the next key bottlenecks in AI server design: efficiently scaling connectivity among larger numbers of GPUs and accelerators inside the server and rack. Marvell said the new switch delivers nearly 2x the lane density of competitive offerings and is designed to provide higher density, lower latency, and improved system efficiency. 

As AI servers integrate more accelerators, traditional PCIe scale-up topologies often require multiple smaller switches to achieve the required connectivity, which can increase power consumption, latency, board complexity, and total cost of ownership. Marvell argues that the higher radix of the Structera S 60260 eliminates the need for multiple devices in many designs, simplifying server architecture while improving resource utilization. The company is also pairing the new switch with its Alaska P PCIe retimer line to offer an end-to-end PCIe connectivity platform, including support for active electrical cables up to seven meters and active optical cable reach beyond seven meters through ecosystem partners. 

The launch is significant because it shows how AI infrastructure bottlenecks are moving deeper into the system stack. Networking for AI is no longer only about front-end Ethernet fabrics or optical interconnects between racks. It is also about how efficiently accelerators communicate within servers, trays, and scale-up domains. Marvell said engineering test samples of the Structera S PCIe 6.0 switch are available now and are featured in a live OFC 2026 demonstration, with customer sampling expected in the third quarter of calendar 2026. The switch is also pin-compatible with Marvell’s newly announced Structera S CXL 3.0 offerings, allowing customers to design a common hardware platform for both PCIe and CXL applications. 

Key points

• Marvell says the Structera S 60260 is the industry’s first 260-lane PCIe 6.0 switch

• Targets AI data center scale-up infrastructure inside accelerator servers and racks

• Designed to reduce switch count, board complexity, latency, and power consumption

• Extends Marvell’s PCIe portfolio following its acquisition of XConn Technologies

• Works with Marvell Alaska P PCIe retimers and ecosystem cable solutions

• Engineering test samples are being shown at OFC 2026

• Customer sampling is expected in Q3 2026

• Pin-compatible with Marvell’s new Structera S CXL 3.0 switch offerings 

“The expansion of AI infrastructure to support next-generation AI workloads continues to move at a breathtaking pace, requiring new, higher-performance scale-up fabrics to eliminate existing bottlenecks,” said Rishi Chugh, vice president and general manager, Data Center Switch Business Unit at Marvell. “XConn has been shipping 256-lane PCIe 5.0 switches since 2022. Now, with the industry’s first 260-lane PCIe 6.0 switch at nearly double the lane density of the competition, Marvell is setting a new standard for PCIe scale-up performance.” 

Tags: #OFC26Marvell
ShareTweetShareSummarizeSummarize
Previous Post

CoreWeave advances AI-native cloud platform with NVIDIA HGX B300

Next Post

Lumentum Showcases 1060nm VCSEL Optical Scale-Up Demo

Jim Carroll

Jim Carroll

Editor and Publisher, Converge! Network Digest, Optical Networks Daily - Covering the full stack of network convergence from Silicon Valley

Related Posts

Semiconductors

Marvell CEO: AI Scaling Has Become a Connectivity Challenge

June 5, 2026
Automotive Networking

Marvell Launches 102.4 Tbps Teralynx T100 Switch

June 1, 2026
Financials

Marvell Posts Record $2.418B Quarter as AI Infrastructure Surges

May 27, 2026
Optical

Marvell Adds Plasmonics to Optical Stack with Polariton Acquisition

April 22, 2026
Video

Marvell: Big Outlook for XPU-Attach

April 16, 2026
Semiconductors

NVIDIA Invests $2B in Marvell to Extend NVLink Fusion AI Ecosystem

March 31, 2026
Next Post

Lumentum Showcases 1060nm VCSEL Optical Scale-Up Demo

Categories

  • 5G / 6G / Wi-Fi
  • AI Infrastructure
  • All
  • Automotive Networking
  • Blueprints
  • Clouds and Carriers
  • Data Centers
  • Enterprise
  • Explainer
  • Feature
  • Financials
  • Last Mile / Middle Mile
  • Legal / Regulatory
  • Optical
  • Quantum
  • Research
  • Security
  • Semiconductors
  • Space
  • Start-ups
  • Subsea
  • Sustainability
  • Video
  • Webinars

Archives

Tags

5G All AT&T Australia AWS Blueprint columns BroadbandWireless Broadcom China Ciena Cisco Data Centers Dell'Oro Ericsson FCC Financial Financials Huawei Infinera Intel Japan Juniper Last Mile Last Mille LTE Mergers and Acquisitions Mobile NFV Nokia Optical Packet Systems PacketVoice People Regulatory Satellite SDN Service Providers Silicon Silicon Valley StandardsWatch Storage TTP UK Verizon Wi-Fi
Converge Digest

A private dossier for networking and telecoms

Follow Us

  • Home
  • About
  • Events Calendar
  • Blueprint Guidelines
  • Privacy Policy
  • Manage Email Delivery
  • NextGenInfra.io

© 2026 Converge Digest - A private dossier for networking and telecoms.

No Result
View All Result
  • Home
  • About
  • Events Calendar
  • Blueprint Guidelines
  • Privacy Policy
  • Manage Email Delivery
  • NextGenInfra.io

© 2026 Converge Digest - A private dossier for networking and telecoms.

This website uses cookies. By continuing to use this website you are giving consent to cookies being used. Visit our Privacy and Cookie Policy.
Go to mobile version