NVIDIA Is Betting Billions on MediaTek
NVIDIA is best known for the GPUs powering much of today’s artificial intelligence boom.
But its latest investment suggests that NVIDIA wants to influence much more than the GPU itself.
On 31 August 2026, NVIDIA and Taiwan-based semiconductor company MediaTek announced an expansion of their long-standing partnership, with NVIDIA investing US$3.5 billion in convertible bonds issued by MediaTek.
The two companies will work together across three major areas: AI infrastructure, local AI computing and automotive technology.
For MediaTek, the partnership provides deeper access to NVIDIA’s AI ecosystem.
For NVIDIA, it creates another way to remain involved even as major technology companies increasingly want to design their own specialised AI processors.
The AI Industry Wants More Custom Chips
For years, NVIDIA GPUs have been the dominant hardware used to train and run advanced AI models.
But companies are increasingly exploring another approach: custom AI chips.
Instead of relying entirely on general-purpose GPUs, hyperscale cloud companies can design processors specifically for their own workloads.
These custom processors are sometimes referred to as XPUs, a broad term covering specialised computing accelerators.
Amazon, Google and other major technology companies are already developing their own silicon as they search for better performance, lower power consumption and more control over their AI infrastructure.
MediaTek wants to become one of the companies helping businesses create those processors.
This Is Where NVLink Fusion Comes In
One of the most important parts of the NVIDIA-MediaTek agreement is NVLink Fusion.
MediaTek will adopt the platform to help customers develop custom AI processors that can integrate directly into NVIDIA’s NVLink-connected computing architecture.
That matters because creating a fast AI chip is only one part of the challenge.
Thousands of processors inside large AI systems also need to communicate with one another extremely quickly.
Technologies such as NVLink provide the high-speed connections that allow GPUs, CPUs and specialised accelerators to work together as a much larger computing system.
With NVLink Fusion, customers could potentially build more customised processors without abandoning NVIDIA’s wider infrastructure ecosystem.
NVIDIA May Be Preparing for a More Diverse AI Chip Market
There is an interesting strategic reason behind this partnership.
The future AI market may not consist entirely of companies purchasing identical NVIDIA GPUs.
Large cloud providers and AI companies increasingly want processors designed specifically around their own applications.
On the surface, that might look like competition for NVIDIA.
But NVIDIA’s strategy appears to be broader.
If custom processors still connect through NVIDIA’s networking, interconnect and software technologies, NVIDIA can remain an important part of the system even when some of the computing is performed by chips designed by another company.
MediaTek’s adoption of NVLink Fusion supports exactly that type of model.
MediaTek Is Moving Beyond Smartphone Chips
MediaTek is already one of the world’s largest fabless semiconductor companies and is widely known for processors used in smartphones, smart TVs and connected devices.
But AI is opening much larger opportunities.
The company has been expanding into high-performance computing and custom silicon, including technology designed for cloud and AI workloads.
The deeper NVIDIA partnership could accelerate that transition.
MediaTek brings expertise in system-on-chip design, connectivity, power efficiency, advanced packaging and custom silicon, while NVIDIA contributes its accelerated computing architecture, GPUs and AI software ecosystem.
Together, the companies are targeting computing from the edge all the way to large-scale AI infrastructure.
The Partnership Also Reaches Your PC
This isn’t only about huge AI systems.
NVIDIA and MediaTek are also collaborating on AI computing closer to the user.
The companies previously worked together on the GB10 Grace Blackwell Superchip used in NVIDIA DGX Spark.
They are now extending that collaboration across multiple generations of RTX Spark and DGX Spark chips, targeting consumer PCs, developer systems and enterprise workstations.
This reflects another trend in AI.
Instead of every AI request being processed somewhere in the cloud, increasingly powerful PCs will be able to run more AI workloads locally.
That can improve responsiveness, reduce cloud dependence and potentially improve privacy for certain applications.
Cars Are Becoming AI Computers Too
The third part of the partnership is automotive technology.
MediaTek and NVIDIA are continuing development of platforms for AI-powered, software-defined vehicles.
Modern cars increasingly use powerful processors for digital cockpits, driver-assistance systems, entertainment, computer vision and other intelligent functions.
MediaTek’s Dimensity Auto platform already integrates NVIDIA technology and can work alongside NVIDIA DRIVE AGX systems.
As vehicles become more software-driven, the amount of computing inside them is expected to continue increasing.
That gives both companies another major market beyond smartphones and traditional computers.
Why This Matters for Asia
This deal also highlights Taiwan’s continued importance in the global semiconductor industry.
MediaTek is headquartered in Taiwan and has grown into one of the world’s largest chip-design companies.
Taiwan is already critical to semiconductor manufacturing through companies such as TSMC.
MediaTek represents another part of that ecosystem: designing sophisticated chips rather than manufacturing them directly.
A US$3.5 billion investment from NVIDIA shows how strategically valuable that expertise has become as the AI industry expands beyond traditional GPUs.
Closing Thoughts
The most interesting part of NVIDIA’s investment in MediaTek isn’t simply the US$3.5 billion figure.
It is what the partnership says about the future of AI hardware.
AI computing is becoming more specialised.
Different companies will increasingly want processors designed specifically for their workloads, power requirements and applications.
Instead of resisting that shift, NVIDIA appears to be positioning its technology as the infrastructure that connects those different chips together.
MediaTek, meanwhile, gets an opportunity to move much deeper into AI infrastructure, PCs and intelligent vehicles.
The AI chip race is therefore becoming much bigger than GPUs alone.