AI Tools & Products News

Nvidia Unveils NVLink Fusion: A New Era of Chip-to-Chip Communication

At the Computex 2025 exhibition in Taipei, Nvidia introduced NVLink Fusion, a new version of its NVLink technology.

This new interconnect system is designed to enhance chip-to-chip communication, enabling more seamless integration of multiple chips in AI systems.

Key Features and Collaborations

NVLink Fusion signifies a strategic transition for Nvidia, from a exclusive “full-stack AI solution” towards a more architectural approach. 

This change allows integration of CPUs and AI chips from other manufacturers into Nvidia-designed server racks.

Companies like Marvell Technology and MediaTek have already announced plans to adopt NVLink Fusion for their custom chip efforts.

NVLink Fusion opens the door for third-party chipmakers to collaborate directly with Nvidia’s ecosystem.

Technical Advancements

Building upon the foundation of NVLink-C2C, NVLink Fusion offers ultra-fast chip interconnect capabilities.

It delivers up to 25 times more energy efficiency and 90 times more area efficiency compared to PCIe Gen 5 PHY on Nvidia chips.

This technology supports high-bandwidth, low-latency communication between processors and accelerators, which is crucial for AI workloads.

Implications of Strategy

By allowing other chipmakers to integrate their technologies with Nvidia’s, the company aims to accelerate the development and deployment of advanced AI systems.

This strategy enables Nvidia to strengthen its foothold data centers and accelerate revenue growth within the rapidly expanding AI and data infrastructure market.

Future Plans

In addition to NVLink Fusion, Nvidia CEO Jensen Huang unveiled a roadmap that includes upcoming releases of Blackwell Ultra chips later this year, Rubin chips, and future Feynman processors set to arrive by 2028.

These developments underscore Nvidia’s transformation from a video game graphics chip maker into a leader in AI chip production.

News Gist

At Computex 2025, Nvidia launched NVLink Fusion, a new interconnect tech enabling third-party AI chip integration.

It boosts chip communication efficiency and fosters collaboration. Nvidia also unveiled DGX Spark AI desktops, with production ramping up in weeks.

Nvidia also launched a desktop version of its AI chips, called DGX Spark, targeting AI researchers and according to Jensen Huang the computer was in full production and would be ready in a “few weeks”.

Leave a Reply

Your email address will not be published. Required fields are marked *

AI Binger
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.