Nvidia announced a $3.5 billion investment in Taiwanese chipmaker MediaTek. As part of the agreement, MediaTek will adopt Nvidia technology to design custom chips intended for artificial intelligence companies and hyperscalers, capable of being directly integrated into data centers based on the American company's platform.
The deal comes at a time when a growing number of AI firms and major cloud providers, such as Amazon, Google, Microsoft, OpenAI, and Anthropic, are allocating resources to develop their own chips. The goal is to rely less on Nvidia's GPUs. Deals like this allow the company led by Jensen Huang to cede ground in custom silicon without losing its dominant position as the central framework of data centers.
“Nvidia is an AI infrastructure company,” said Dion Harris, senior director of HPC and AI infrastructure solutions for hyperscalers at Nvidia, during a conference with journalists. “We expanded beyond pure compute chips years ago.”
The partnership reflects the circular nature of Nvidia's financial operations in recent years: the company often invests in firms that end up reinforcing its own ecosystem. MediaTek will gain access to the NVLink Fusion ecosystem, which includes NVLink technology. This allows any chip, even if it is not from Nvidia, to communicate at high speed with the rest of the infrastructure.

A strategy to remain the standard
Last week, Nvidia announced a similar collaboration with Amazon Web Services, although without direct investment. AWS will deploy an additional two million Nvidia GPUs in its infrastructure and will also integrate NVLink Fusion. With the current agreement, MediaTek will be able to continue manufacturing custom chips designed by cloud companies or AI labs, tailored to their specific workloads, and rely on Nvidia's technology stack and scale ecosystem.
“Basically, every cloud and every model creator is deploying our platform in some way,” Harris explained. “By being able to offer this extension to their customers, MediaTek allows them to standardize rack-scale infrastructure in their AI factories and deploy their custom chips alongside others using the same standard platform.”
Harris emphasized that this is about opening the ecosystem to MediaTek's entire customer base. The Taiwanese company has not disclosed public details about its portfolio of custom AI chip clients, but it is clear that Nvidia aims to boost it in that segment whenever it can support growth.
MediaTek has been gradually consolidating its ASIC business for data centers. In June, it estimated that this area would generate $2 billion in revenue by 2026 and aimed to capture a larger share of the market in the coming years. Its experience in custom chips for smartphones, smart homes, cars, and wireless communications provides a solid foundation for this leap.
Beyond data centers
In addition to the hyperscaler front, both companies will continue collaborating on the DGX Spark, the small desktop computer aimed at AI developers. They also extended their joint work on the RTX Spark, Nvidia's bet to bring its AI technology to consumer PCs. In the automotive sector, they will continue developing platforms for software-defined vehicles powered by AI in the era of “physical AI.”
MediaTek's automotive platforms already use Nvidia's RTX graphics for smart cabin systems and work alongside Nvidia Drive AGX, the onboard computing platform intended for autonomous driving workloads.
“AI is transforming every computing platform, from the largest AI factories in the world to PCs and cars,” stated Jensen Huang, founder and CEO of Nvidia. “Together we are building platforms that bring NVIDIA's accelerated computing to new markets and give customers the freedom to create differentiated AI systems at a massive scale.”
With this strategic investment, Nvidia not only strengthens its position against the advance of proprietary chips from major tech companies but also seeks to remain the benchmark standard upon which the AI infrastructures of the future are built. The bet on MediaTek allows it to accompany customization without relinquishing control of the ecosystem that remains the center of gravity of the industry.