TECH NEWS

AI networking startups race to replace Nvidia's NVLink

As Nvidia expands its influence through its NVLink Fusion tech, rival networking vendors are scrambling to bring alternative interconnects and switches to market.

At the AI Infra Summit this week, Delos Data and Cornelis Networks officially entered the scale up networking race.

Scale-up fabrics, like NVLink, are what have allowed Nvidia to make eight, 72, and now 576 GPUs behave as one enormous AI accelerator. To catch up, rivals like AMD have embraced emerging protocols like Ultra Accelerator Link.

Today, these protocols are largely being tunneled over standard Ethernet switches. For instance, AMD is using Broadcom’s 102.4 Tbps Tomahawk 6-based switches connecting to custom I/O dies on the MI455X.

Purpose-built UALink switches and physical interconnects remain elusive, but that won’t be the case for long if Cornelis and Delos have their way.

The two companies are approaching this challenge from a few different angles, including standardization, software optimization,...

Copyright of this story solely belongs to www.theregister.com. To see the full text click HERE

https://cdn1.expresscomputer.in/wp-content/uploads/2026/04/05215204/Artificial-Intelligence.png

Managing the Algorithm: Who is accountable when AI makes a business decision?

By Dr Kavita Kanabar, Associate Professor, HR and General Management, Vivekanand Education Society’s Institute of Management Studies and Research (VESIMSR) A loan applicant is turned down by a system, not a person. A résumé is filtered out before a recruiter ever opens it. A delivery app quietly raises its

https://cdn1.expresscomputer.in/wp-content/uploads/2026/07/23085230/AI-Security-AI-Attack.jpg

In AI security, the best defence is a great offense

By Chaim Mazal, CISO, GitLab India recorded 265.52 million threat detections across 8 million endpoints in 2025, and the number keeps climbing despite increased security investments. At the same time, the level of expertise an attacker needs to run a sophisticated attack continues to fall. Advanced AI models allow

https://substackcdn.com/image/fetch/$s_!xGnO!,w_1200,h_675,c_fill,f_jpg,q_auto:good,fl_progressive:steep,g_auto/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9d660...

Analysis: only ~2.3 GW of planned data center capacity is delayed by direct project-level policies like local moratoriums and New York's data center EO

Sponsor Posts Subquadratic: the LLM built for 12M-token reasoning — SubQ can reason across entire codebases and document sets in one pass with no RAG workarounds. Read how SubQ 1.1 Small holds near-perfect retrieval out to 12M tokens. Cloudflare Connect 2026: The Agentic Conference — Stop preparing for agents. Start running

https://hackster.imgix.net/uploads/attachments/2001938/_FI7DW1ymWo.blob?auto=compress%2Cformat&w=600&h=450&fit=min

A Cool New Way to Make Custom LED Display

You’ve probably noticed that many consumer electronics devices have LED displays — not LED matrix displays or OLED displays, but displays with various icons and indicators illuminated by LEDs. They look nice and they’re very readable at a glance, but you almost never see them in custom builds. That’

Read more

Managing the Algorithm: Who is accountable when AI makes a business decision?

In AI security, the best defence is a great offense

Analysis: only ~2.3 GW of planned data center capacity is delayed by direct project-level policies like local moratoriums and New York's data center EO

A Cool New Way to Make Custom LED Display