Nvidia is extending its AI infrastructure strategy beyond its traditional GPU franchise with a $3.5 billion investment in MediaTek, underscoring the chipmaker’s growing interest in custom silicon as artificial intelligence workloads diversify. The transaction, which represents roughly 3.6% of Nvidia’s recent $96.2 billion quarterly revenue, gives Nvidia a deeper financial and technological relationship with one of Asia’s largest chip designers while its biggest customers increasingly explore alternatives to general-purpose GPUs.

Nvidia Takes a Major Stake in MediaTek’s AI Expansion

Nvidia agreed to purchase approximately $3.5 billion of MediaTek’s convertible bonds, representing about 90% of the Taiwanese company’s $3.9 billion offshore offering. Alphabet also participated, although the companies did not disclose the size of its investment.

The financing accompanies a broader technology partnership. MediaTek will adopt Nvidia’s NVLink Fusion platform, allowing custom-designed processors to connect directly with Nvidia’s rack-scale computing infrastructure. The companies are targeting AI infrastructure, edge AI and software-defined vehicles, extending their cooperation across several high-growth technology markets.

The relationship also includes RTX Spark and DGX Spark systems, where MediaTek’s system-on-chip capabilities are combined with Nvidia GPUs for AI-focused PCs and developer workstations.

Why Custom AI Chips Are Becoming More Important

MediaTek is building a significant business around application-specific integrated circuits designed for targeted AI workloads. The company expects its data-center ASIC operation to generate more than $2 billion in revenue this year and estimates that its addressable market could reach $80 billion by 2027.

Its ambitions are increasing as well. MediaTek is now targeting a 15% to 20% share of that market, compared with an earlier objective of 10% to 15%. Its first AI accelerator ASIC is scheduled for mass production in the fourth quarter of 2026, with a second-generation processor expected to enter high-volume production in 2028.

That timing is important because major AI companies are pursuing customized silicon to improve performance, efficiency and control over increasingly specialized workloads.

Could Nvidia Benefit From Financing Its Own Ecosystem?

Nvidia’s investment creates an unusual strategic dynamic. The convertible bonds carry no regular interest, meaning Nvidia can potentially benefit from equity appreciation if MediaTek’s shares rise above the conversion threshold. Nvidia therefore gains financial exposure to MediaTek while strengthening a technology partnership that could generate additional demand for Nvidia’s networking and computing infrastructure.

Chief Executive Jensen Huang has rejected suggestions that the arrangement represents circular financing, arguing that the companies remain separate businesses. Nevertheless, the structure highlights how Nvidia is positioning itself within the broader AI-chip ecosystem rather than relying exclusively on GPU sales.

Google’s involvement and MediaTek’s work on custom accelerators add another layer of significance. As companies including Google and OpenAI develop specialized AI processors, Nvidia’s partnership with a major ASIC designer could provide an avenue to participate in the custom-silicon transition rather than simply compete against it.

For investors, the key question is whether Nvidia can maintain its GPU leadership while successfully capturing value from the expanding ecosystem around custom chips, networking and AI infrastructure. MediaTek’s ASIC production milestones, customer adoption and market-share progress will be important indicators of whether the $3.5 billion investment becomes a strategic advantage as AI computing continues to evolve.