Nvidia Backs MediaTek With $3.5B to Keep Custom AI Silicon Inside Its Orbit
The multibillion-dollar investment allows Taiwanese chip designer MediaTek to build custom hyperscaler silicon on Nvidia's NVLink interconnects.
Key highlights · 2 min read
- Nvidia is investing $3.5 billion into Taiwanese semiconductor designer MediaTek, a strategic maneuver aimed at ensuring next-generation custom artificial intelligence chips remain tethered to Nvidi…
- Under the agreement, MediaTek will incorporate Nvidia's NVLink Fusion technology into custom application-specific integrated circuits (ASICs) built for cloud hyperscalers and frontier AI labs, as f…
- The move comes as tech giants such as Microsoft, Alphabet's Google, Amazon, OpenAI, and Anthropic accelerate internal chip initiatives to cut their dependence on pricey Nvidia GPUs.
The Scale ReportNvidia is investing $3.5 billion into Taiwanese semiconductor designer MediaTek, a strategic maneuver aimed at ensuring next-generation custom artificial intelligence chips remain tethered to Nvidia's proprietary data center architecture.
Under the agreement, MediaTek will incorporate Nvidia's NVLink Fusion technology into custom application-specific integrated circuits (ASICs) built for cloud hyperscalers and frontier AI labs, as first reported by TechCrunch. The interconnect standard allows third-party processors to link directly into Nvidia-powered server racks and communicate seamlessly alongside standard graphics processors.
Shifting the Data Center Moat
The move comes as tech giants such as Microsoft, Alphabet's Google, Amazon, OpenAI, and Anthropic accelerate internal chip initiatives to cut their dependence on pricey Nvidia GPUs. By embedding NVLink directly into custom silicon built by MediaTek, Nvidia is positioning its rack-scale architecture as the default computing fabric, capturing value from the broader infrastructure even when customers design their own silicon.
"Nvidia is an AI infrastructure company," Dion Harris, senior director of HPC and AI hyperscaler infrastructure solutions at Nvidia, said during a media briefing. Harris emphasized that the integration allows MediaTek clients to standardize on Nvidia's rack-scale layout across their compute clusters, deploying custom accelerators alongside standard GPUs using a single platform.
The capital infusion also accelerates MediaTek's push into enterprise hardware. While historically recognized for smartphone and consumer electronics silicon, MediaTek stated in June that its custom data center ASIC division is projected to generate $2 billion in revenue in 2026.
Expanding Beyond the Server Room
The collaboration extends across edge computing, personal computers, and automotive platforms. The two semiconductor firms will continue work on DGX Spark developer desktop units while expanding into consumer AI PCs with the RTX Spark initiative. In transportation, MediaTek will continue pairing its smart cockpit platforms with Nvidia's RTX graphics and the Nvidia Drive AGX autonomous driving system.
"AI is transforming every computing platform, from the world’s largest AI factories to the PC and the car," Jensen Huang, founder and chief executive of Nvidia, said in a statement detailing the multi-market expansion.
Reporting based on coverage from AI News & Artificial Intelligence | TechCrunch.




