Skip to content
Breaking:

Nvidia Purchases $3.5B in MediaTek Convertible Debt to Deepen Chip Partnership

The deal expands technical collaboration between the two semiconductor makers across enterprise AI accelerators, personal computers, and automotive silicon.

By The Company Wire3 min read
Share
Nvidia — Nvidia Purchases $3.5B in MediaTek Convertible Debt to Deepen Chip Partnership
Nvidia — Nvidia Purchases $3.5B in MediaTek Convertible Debt to Deepen Chip Partnership. Photo: SiliconANGLE.

Nvidia Corp. has purchased $3.5 billion worth of convertible bonds from MediaTek Inc. as part of a major expansion of the two semiconductor makers' strategic partnership, as first reported by SiliconANGLE. The broad agreement encompasses custom artificial intelligence hardware for enterprise data centers, personal computing platforms, and automotive silicon.

The bond structure permits Nvidia to convert its debt holding into equity at a future date, following a similar move earlier this year when Nvidia acquired a $2 billion stake in competing chip designer Marvell Technologies Inc. Headquartered in Taiwan, MediaTek is widely recognized for supplying smartphone and connected device processors, but has rapidly expanded a custom silicon division that designs specialized AI accelerators for business clients.

That custom design unit serves as the immediate focus of the expanded collaboration. Enterprise customers that hire MediaTek to design tailor-made AI accelerators will now have the option to incorporate Nvidia’s NVLink Fusion networking technology directly onto their chip layouts. The interconnect technology enables high-speed data exchanges between graphics processing units and central processing units across server racks, achieving transfer speeds more than 10 times faster than a standard PCIe 6.0 connection.

The integration builds upon a technical enhancement Nvidia introduced last week supporting a memory architecture called NVHBM. By moving the memory controller from the GPU logic die directly onto the High Bandwidth Memory die, the approach boosts data movement performance while reducing total power consumption.

Beyond data center infrastructure, the expanded partnership focuses on personal computing hardware. Following the 2025 release of Nvidia's DGX Spark developer workstation and the June debut of the RTX Spark processor for laptops and desktops—both engineered with MediaTek's assistance—the companies plan to co-develop multiple future generations of both product lines.

Automotive applications mark the third core pillar of the updated relationship. MediaTek’s Dimensity Auto line of system-on-chips, which powers digital display systems inside vehicles, currently includes an Nvidia graphics rendering chiplet. Under the expanded agreement, both firms plan to continue joint engineering work on processors designed for software-defined vehicles.

“Together, we’re building platforms that bring Nvidia accelerated computing to new markets and give customers the freedom to create differentiated AI systems at enormous scale,” Nvidia Chief Executive Officer Jensen Huang said in a statement. The alliance comes as MediaTek recently doubled its full-year revenue forecast for its data center chip design group to $2 billion, with plans to expand that division severalfold in the years ahead.

Sources

  1. SiliconANGLE

Company: Nvidia

Written by

The Company Wire

Newsroom · San Francisco

Inside the companies building what’s next. Reporting on startups, technology, funding and the people shaping them.