Nvidia’s NVLink Fusion & its Impact on the AI Market

Share this article
Share this article
Prioritise Us on Google
Nvidia’s NVLink Fusion is designed to let industries build semi-custom AI infrastructure | Credit: Nvidia
Facing shifting AI spending from cloud providers, Nvidia launches NVLink Fusion that enables integration of custom processors into AI infrastructure

The AI boom that has propelled Nvidia to become the world’s most valuable chip company now faces a critical test as spending patterns shift across the industry.

After two years of unprecedented growth driven by the rush to build AI infrastructure, cloud computing giants and governments are beginning to reassess their investment strategies, creating uncertainty for the semiconductor sector’s leading player.

The initial wave of AI infrastructure development saw companies and nations commit hundreds of billions of dollars to data centres equipped with Nvidia’s specialised processors. 

However, as these systems come online and demonstrate their capabilities, buyers are becoming more selective about future purchases.

Major cloud providers including Microsoft and Google, which operate vast networks of data centres, have signalled plans to moderate their AI hardware spending.

As a result, Nvidia has unveiled technology that allows companies to plug custom chips into its AI infrastructure – a strategic shift as the semiconductor company faces potential slowdowns in AI spending from major cloud computing providers.

What is Nvidia’s NVLink Fusion?

The technology, called NVLink Fusion, is CEO Jensen Huang’s answer to growing concerns about the sustainability of massive sovereign AI infrastructure deals – and signals from companies such as Microsoft and Alphabet’s Google about cuts to AI spending.

Key facts:
  • Nvidia is striving to expand AI market grip through NVLink Fusion
  • Nvidia is facing potential AI spending cuts from leading cloud firms
  • Taiwan’s tech ecosystem is crucial for Nvidia's growth

NVLink Fusion functions as a platform that enables companies to integrate their own specialised processors into Nvidia’s AI systems.

The approach transforms Nvidia from purely a chip supplier into a platform provider, potentially expanding its customer base beyond the current focus on large-scale data centre projects.

“Instead of having to build the entire rack of equipment themselves, (companies) could innovate or differentiate on the custom (chip) itself,” Nick Kucharewski, VP at Marvell Technology, tells Reuters.

The platform strategy aims to drive demand for Nvidia’s underlying AI network and data centre components by attracting hardware developers to build products that utilise the Fusion system.

Nvidia targeting enterprise market with new server line

Jensen, who spent a week in Taiwan at the Computex trade fair, also announced a line of servers described as an “enterprise AI supercomputer” to target business customers beyond the current focus on large technology companies and governments.

CEO of Nvidia, Jensen Huang

These enterprise servers are designed to handle multiple computing tasks including graphics processing, virtual machines and AI applications.

Jensen pitched the technology as opening access to a multi-billion dollar market where customers can use the systems for a range of functions.

However, the enterprise market presents different challenges compared to Nvidia's current customer base.

Deals tend to be smaller in scale than sovereign data centre projects and require more time and resources to secure.

Furthermore, US restrictions on high-technology exports have forced Nvidia to withdraw certain chips from China and develop alternative products that comply with evolving government policy, resulting in a loss of market share in that region.

Meanwhile, Microsoft and Google have indicated plans to reduce AI infrastructure spending. These companies have been among Nvidia’s largest customers for AI chips used in data centres.

Taiwan supply chain being crucial for AI infrastructure expansion

While Jensen announced deals worth hundreds of billions of dollars in regions including the Gulf states during recent travels, analysts suggest such large-scale agreements will become less frequent.

Youtube Placeholder

“Is every country going to announce a US$10bn or US$50bn data centre like the Saudis? Of course not,” Seaport Research Analyst Jay Goldberg tells Reuters.

“They’re sort of running out of obvious deals.”

When asked about plans to address potential AI spending slowdowns, Jensen responds: “AI infrastructure is being built out (everywhere) - that’s one of the reasons I'm travelling around the world... AI infrastructure is going to be a part of society.”


Explore the latest edition of AI Magazine and be part of the conversation at our global conference series, Tech & AI LIVE

Discover all our upcoming events and secure your tickets today.

Also sign up to our free weekly newsletter for the latest insights and stories straight into your inbox.


AI Magazine is a BizClik brand 

Company portals