NVIDIA and MediaTek Deepen AI Partnership With $3.5 Billion Investment

NVIDIA and MediaTek are expanding their long-running partnership across AI infrastructure, local AI computing and automotive technology, in a move that could reshape how custom AI processors connect to large-scale NVIDIA systems.

Under the expanded collaboration, MediaTek will adopt NVIDIA’s NVLink Fusion platform, giving hyperscalers, cloud providers and frontier AI developers a prevalidated route for building custom AI accelerators, known as XPUs, that can connect to NVIDIA’s rack-scale AI infrastructure.

NVIDIA is also investing $3.5 billion in convertible bonds issued by MediaTek, further strengthening the relationship between the two semiconductor companies.

The partnership combines NVIDIA’s expertise in accelerated computing, AI software, graphics and interconnect technology with MediaTek’s capabilities in custom silicon, system-on-chip design, connectivity, advanced packaging and power-efficient computing.

NVIDIA and MediaTek Expand Across Three AI Markets

The companies say their expanded collaboration will focus on three major areas.

AI infrastructure: MediaTek will work within the NVIDIA NVLink Fusion ecosystem to help customers develop custom AI infrastructure capable of integrating with NVIDIA rack-scale systems and AI factories.

Local AI computing: The companies will continue developing generations of NVIDIA RTX Spark and DGX Spark PC chips. These systems combine NVIDIA GPU technology with MediaTek SoCs for consumer PCs, AI development systems and enterprise workstations.

Automotive: NVIDIA and MediaTek will continue developing platforms for software-defined vehicles powered by increasingly sophisticated AI systems.

NVIDIA CEO Jensen Huang said AI is increasingly transforming computing across data centers, PCs and vehicles.

“AI is transforming every computing platform — from the world’s largest AI factories to the PC and the car,” Huang said.

He described MediaTek as having strong expertise in SoC design, connectivity, performance and power efficiency, saying the partnership could bring NVIDIA accelerated computing into additional markets.

MediaTek Vice Chairman and CEO Rick Tsai said NVIDIA’s investment strengthens a relationship spanning cloud infrastructure, local computing and automotive applications.

What Is NVIDIA NVLink Fusion?

A major part of the announcement is MediaTek’s adoption of NVIDIA NVLink Fusion, a platform designed to help companies build custom AI accelerators that can operate within NVIDIA-connected rack-scale systems.

Instead of developing every component surrounding a custom accelerator independently, customers can use NVIDIA technologies for key elements such as connectivity, memory architecture and rack-scale integration.

The approach allows companies to concentrate more of their engineering resources on the computing architecture that differentiates their AI systems.

The NVLink Fusion platform includes several technologies designed for custom XPU development.

Among them is the NVIDIA NVLink Fusion chiplet, which connects custom XPUs to NVIDIA’s NVLink scale-up fabric through photonics or electrical interconnects.

The platform also includes NVLink-C2C, which provides high-bandwidth connectivity between XPUs, NVIDIA CPUs and other compatible processors.

Another component is NVIDIA NVHBM, which is designed to support customized memory configurations aimed at improving bandwidth and energy efficiency.

Why Custom AI Accelerators Matter

AI companies and cloud providers are increasingly looking beyond traditional off-the-shelf processors as workloads become larger and more specialized.

Building a custom accelerator, however, is extremely complicated. Developers have to deal with advanced packaging, HBM memory, high-speed SerDes, I/O, networking, manufacturing and system-level qualification.

NVIDIA and MediaTek’s expanded collaboration is intended to simplify that process.

Customers can bring their XPU designs to MediaTek and customize areas such as connectivity, memory, packaging, performance and power characteristics around their workloads.

At the same time, NVIDIA provides access to its NVLink connectivity and broader rack-scale infrastructure technologies.

The companies say this can reduce development complexity and potentially accelerate the deployment of semi-custom AI factories.

MediaTek Is Also Working With NVIDIA on Local AI PCs

The partnership extends beyond massive data centers.

MediaTek previously collaborated with NVIDIA on the GB10 Grace Blackwell Superchip, which powers NVIDIA DGX Spark.

The system combines an NVIDIA Blackwell GPU with a Grace CPU using NVLink-C2C connectivity, bringing significant AI computing capability to a compact local system.

The companies are now extending their work into the next generation of AI-focused PCs through NVIDIA RTX Spark.

That effort reflects the growing push toward running generative and agentic AI workloads locally rather than relying entirely on cloud infrastructure.

For consumers and businesses, local AI computing can potentially provide faster access to AI applications while reducing reliance on remote processing for certain workloads.

NVIDIA and MediaTek Push Deeper Into AI Cars

Automotive technology is the third major pillar of the expanded relationship.

MediaTek and NVIDIA are working on multiple generations of platforms for AI-powered software-defined vehicles, combining MediaTek’s automotive SoC capabilities with NVIDIA’s accelerated computing, graphics and AI software technologies.

MediaTek’s Dimensity Auto platforms integrate NVIDIA technologies for intelligent vehicle cockpits, including advanced AI and NVIDIA RTX graphics.

The platforms can also work alongside NVIDIA DRIVE AGX technology.

The companies intend to build on this architecture in future generations as vehicles become increasingly software-defined and AI-driven.

What the $3.5 Billion NVIDIA Investment Means

NVIDIA’s $3.5 billion investment in MediaTek convertible bonds adds a significant financial dimension to the technology partnership.

The investment aligns with NVIDIA’s broader strategy of expanding its computing ecosystem beyond its own processors and systems.

For MediaTek, the relationship provides a deeper connection to NVIDIA’s AI infrastructure and accelerated-computing ecosystem while allowing the company to apply its custom silicon and SoC expertise across cloud, PC and automotive markets.

The announcement does not mean MediaTek is replacing NVIDIA’s GPUs. Instead, the companies are building an ecosystem in which MediaTek technology can work alongside NVIDIA’s platforms across different types of AI computing.

What Happens Next?

The most important development to watch will be how quickly customers adopt NVLink Fusion for custom AI accelerator projects.

The success of the collaboration will ultimately depend on whether hyperscalers, cloud providers and AI developers use the platform to bring specialized XPUs into production-scale NVIDIA-connected infrastructure.

At the same time, NVIDIA and MediaTek are expected to continue advancing their work on AI PCs and software-defined vehicles.

Together, the three areas show how the AI computing market is expanding beyond traditional data-center GPUs — toward custom accelerators, local AI systems and intelligent vehicles.