menu
close

Nvidia Opens AI Chip Ecosystem with NVLink Fusion Technology

Nvidia has unveiled NVLink Fusion, a groundbreaking technology that allows chip-to-chip communication between Nvidia GPUs and third-party processors. This strategic move enables companies to build semi-custom AI infrastructure using Nvidia's computing fabric while incorporating their own custom chips. Major tech firms including MediaTek, Marvell, Qualcomm, and Fujitsu have already committed to adopting the technology for their AI development efforts.
Nvidia Opens AI Chip Ecosystem with NVLink Fusion Technology

In a significant shift from its traditionally closed ecosystem approach, Nvidia has announced NVLink Fusion, new silicon technology that opens its high-speed interconnect capabilities to third-party chip manufacturers.

Revealed at Computex 2025 in Taiwan, NVLink Fusion allows companies to integrate non-Nvidia CPUs and accelerators with Nvidia's GPUs in AI data centers. This marks the first time Nvidia has permitted its proprietary NVLink technology—previously restricted to Nvidia-only systems—to connect with processors from other manufacturers.

"A tectonic shift is underway: for the first time in decades, data centers must be fundamentally rearchitected — AI is being fused into every computing platform," said Jensen Huang, Nvidia's founder and CEO, during his keynote presentation.

The technology addresses a critical bottleneck in AI development: the need for ultra-fast data exchange between multiple chips in large-scale AI systems. NVLink Fusion delivers up to 25 times more energy efficiency and 90 times more area efficiency than standard PCIe connections, enabling the creation of more powerful and efficient AI factories.

Early adopters include MediaTek, Marvell, Alchip Technologies, Astera Labs, Synopsys, and Cadence, who will use NVLink Fusion to develop custom silicon for AI applications. Additionally, Fujitsu and Qualcomm plan to integrate their custom CPUs with Nvidia GPUs using the technology.

This strategic move strengthens Nvidia's position in the rapidly evolving AI chip market, where it currently holds between 70-95% market share for AI accelerators. By opening its ecosystem while maintaining control of the critical interconnect technology, Nvidia aims to remain at the center of AI infrastructure development even as cloud providers like Google, Microsoft, and Amazon develop their own custom processors.

NVLink Fusion represents Nvidia's adaptation to changing market dynamics, as the company transitions from its roots as a gaming GPU manufacturer to becoming the dominant provider of chips powering the AI industry since ChatGPT's launch in 2022.

Source: Reuters

Latest News