Quick Summary
- Nvidia is investing $3.5 billion in Taiwanese chipmaker MediaTek through convertible bonds.
- The deal gives MediaTek access to NVLink Fusion, Nvidia’s interconnect technology for custom AI chips.
- This move is central to Nvidia’s AI chip strategy as major cloud companies build their own silicon.
- Amazon, Google, Microsoft, OpenAI, and Anthropic are all developing custom chips to cut Nvidia costs.
- MediaTek expects its data center chip business to bring in $2 billion in revenue this year.
- The partnership also covers desktop AI computers and software-defined vehicles.
- Nvidia is using its balance sheet to stay embedded in AI infrastructure even as competition grows.
Nvidia AI chip strategy just took a new turn. The company is putting $3.5 billion into Taiwanese chipmaker MediaTek, a deal that lets MediaTek build custom AI chips that plug directly into Nvidia’s data center systems. On the surface, this looks like a simple investment. Underneath, it shows how Nvidia plans to stay essential to AI infrastructure even as its biggest customers try to build around it.
The announcement came on August 31, 2026. Nvidia will buy convertible bonds issued by MediaTek, a financial structure that lets Nvidia act as a lender now with the option to become a shareholder later. It is Nvidia’s largest direct investment outside the United States. The deal also extends an existing partnership that already covers personal computers and automotive technology.
Inside the Nvidia and MediaTek Agreement
Under the agreement, MediaTek will adopt Nvidia’s NVLink Fusion technology. This gives MediaTek the tools to design custom chips for AI companies and cloud providers, chips that can then connect smoothly to Nvidia-based data centers. Instead of building every part of a chip system from scratch, MediaTek’s customers can rely on Nvidia’s existing scale-up and scale-out infrastructure.
Dion Harris, Nvidia’s senior director of HPC and AI hyperscaler infrastructure solutions, described the company’s role plainly on a call with reporters. He framed Nvidia as an AI infrastructure company rather than a simple chip vendor, noting that it moved beyond pure computing hardware years ago. That framing positions Nvidia as the backbone that other companies build on top of, rather than just a GPU supplier.
MediaTek has not shared full details about which companies it will serve with these custom chips. What is known is that MediaTek has been steadily growing its data center chip business, known as ASIC design, which stands for application-specific integrated circuits. These are chips built for one specific job rather than general-purpose computing.
Nvidia AI Chip Strategy Follows a Familiar Pattern
Nvidia has spent the past few years investing directly in companies that circle back into its own ecosystem. This deal follows that same pattern. By funding MediaTek’s custom chip ambitions, Nvidia gains a foothold in a market segment it does not fully control on its own, since many hyperscalers now want chips tailored to their exact workloads instead of general-purpose GPUs.
This is not Nvidia stepping back from competition. It is Nvidia making sure that even chips built to reduce reliance on its GPUs still depend on its surrounding technology. NVLink Fusion is the clearest example. It allows non-Nvidia chips to communicate quickly using Nvidia’s interconnect standard, which means custom silicon still runs inside a Nvidia-shaped infrastructure.
Just one week before the MediaTek announcement, Nvidia struck a related deal with Amazon Web Services. That partnership did not include a direct investment, but AWS agreed to deploy an additional two million Nvidia GPUs and integrate NVLink Fusion into its systems. Together, these two deals point to a consistent approach: support the shift toward custom chips while keeping Nvidia’s technology at the center of it.
Why Big Tech Keeps Building Its Own AI Chips
The MediaTek deal did not happen in isolation. Amazon, Google, Microsoft, OpenAI, and Anthropic have all been developing their own AI chips to reduce how much they depend on Nvidia’s GPUs. Cost is the main driver here. Building custom silicon lets these companies design hardware around their specific AI models, which can lower long-term expenses compared to buying general-purpose chips at scale.
According to reports, MediaTek’s custom AI chip business is expected to bring in around $2 billion in revenue this year, with the total addressable market for custom AI chips potentially reaching $80 billion by 2027. That scale explains why Nvidia wants a stake in how this part of the industry grows, rather than watching from the sidelines.
It is also noted that the investment marks Nvidia’s largest direct financial commitment outside the United States to date. That detail underscores how seriously Nvidia is treating the custom silicon trend, even as it continues to dominate general-purpose AI computing chip sales worldwide.
How MediaTek’s Customers Benefit from NVLink Fusion
NVLink Fusion is essentially a bridge. It lets chips that are not made by Nvidia still communicate at high speed with systems that are. For a company designing its own AI chip, this removes a major engineering burden. Instead of building custom connectors and communication protocols, that company can rely on infrastructure Nvidia has already tested at scale.
Most major clouds and AI labs already run some version of Nvidia’s platform in their data centers. By offering MediaTek’s customers a way into that same standard, Nvidia lets them build custom chips without walking away from the rack-scale approach the rest of the industry already relies on.
This matters for smaller or mid-sized AI companies too. Not every organization has the resources to build a full data center architecture from the ground up. By working through MediaTek and Nvidia’s combined technology, these companies can access proven infrastructure without needing to replicate it themselves.
Nvidia and MediaTek Are Also Teaming Up on PCs and Cars
The Nvidia and MediaTek relationship goes beyond cloud infrastructure. The two companies will keep collaborating on DGX Spark, Nvidia’s small desktop AI computer built for developers. They are also extending work on RTX Spark, which brings Nvidia’s AI technology into consumer AI PCs.
Automotive technology is another shared focus. The companies plan to keep developing platforms for AI-powered, software-defined vehicles. MediaTek’s automotive chips already work with Nvidia’s RTX graphics for vehicle cockpit displays, alongside Nvidia Drive AGX, which powers autonomous driving systems.
This shows that Nvidia’s AI chip strategy is not limited to massive data centers. It also reaches into everyday consumer devices and cars.
The Road Ahead for Nvidia and MediaTek
Nvidia CEO Jensen Huang summarized the philosophy behind the deal in a public statement. He described AI as reshaping every type of computing platform, from massive data centers down to personal computers and cars. He also said the partnership is meant to bring Nvidia’s accelerated computing into new markets while giving customers room to build differentiated AI systems at large scale.
That freedom comes with a condition. Companies building their own AI chips still need infrastructure to connect them, cool them, and scale them across data centers. Nvidia’s bet is that even in a future with more diverse chip designs, its interconnect and infrastructure technology remains the common thread tying the industry together.
For MediaTek, the benefits are more direct. The investment strengthens its balance sheet and validates its push into data center chips, an area where it has historically trailed larger rivals. It also positions MediaTek to compete for a larger share of a market that could be worth tens of billions of dollars within a few years.
Final Thoughts
Nvidia’s $3.5 billion investment in MediaTek is more than a single funding announcement. It reflects a deliberate strategy to remain central to AI infrastructure, even as major tech companies build custom chips specifically to reduce their reliance on Nvidia’s GPUs. By backing MediaTek’s push into data center silicon and folding it into the NVLink Fusion ecosystem, Nvidia positions itself to benefit from the custom chip trend rather than lose ground to it.
This approach extends across cloud computing, consumer AI devices, and automotive technology, showing how broadly Nvidia’s ambitions now stretch. As more companies pursue their own AI chips, deals like this one suggest that Nvidia’s infrastructure, rather than its GPUs alone, may be what keeps it at the center of the AI industry for years to come.
Discover how AI is reshaping technology, business, and healthcare—without the hype.
Visit InfluenceOfAI.com for easy-to-understand insights, expert analysis, and real-world applications of artificial intelligence. From the latest tools to emerging trends, we help you navigate the AI landscape with clarity and confidence
