NVIDIA and MediaTek expand collaboration across AI and autos

NVIDIA and MediaTek announced an expanded collaboration covering AI infrastructure, local AI computing and automotive applications.

As part of the collaboration, MediaTek will adopt the NVIDIA NVLink Fusion platform to provide hyperscalers, cloud service providers and AI model developers with a platform for developing custom XPUs that can connect to NVIDIA NVLink-based rack-scale AI systems.

The collaboration combines NVIDIA’s accelerated computing, AI, graphics and software platforms with MediaTek’s capabilities in custom silicon, high-performance computing, power-efficient system-on-chip (SoC) design, advanced packaging, interconnects and connectivity.

NVIDIA has also agreed to invest $3.5 billion in convertible bonds issued by MediaTek.

NVIDIA and MediaTek are collaborating in three areas:

• AI infrastructure: MediaTek will work with the NVIDIA NVLink Fusion ecosystem to support customers developing custom AI infrastructure designed to integrate with NVIDIA rack-scale systems.
• Local AI computing: The companies will continue developing multiple generations of NVIDIA RTX Spark and DGX Spark PC chips for consumer PCs, AI development systems and enterprise workstations. The platforms combine NVIDIA GPUs with MediaTek SoCs.
• Automotive: MediaTek and NVIDIA will continue developing platforms for AI-enabled, software-defined vehicles and physical AI applications.

Building custom AI infrastructure with NVLink Fusion

MediaTek will offer the NVIDIA NVLink Fusion platform as a design foundation for customers developing custom AI accelerators. The platform is intended to support integration with future NVIDIA architectures.

NVLink Fusion provides a system-validated foundation for multi-die XPU development, covering areas from silicon and advanced packaging to rack-scale systems.

The platform includes several technologies for custom XPU development:

• NVIDIA NVLink Fusion chiplet: Connects XPUs to the NVIDIA NVLink scale-up fabric through NVIDIA photonics or electrical interconnects.
• NVIDIA NVLink-C2C: Provides high-bandwidth, power-efficient connectivity between XPUs, NVIDIA Vera CPUs and other compatible processors.
• NVIDIA NVHBM: Supports customized memory configurations intended to increase memory bandwidth and power efficiency while allocating more silicon area to compute.

Developing a custom accelerator is one part of deploying XPUs in rack-scale AI systems. Multi-die architectures, advanced packaging, high-speed SerDes, HBM, I/O and scale-up networking also require engineering, qualification and supply-chain support.

The collaboration allows customers to focus on their custom compute designs while NVIDIA and MediaTek provide technologies and support for NVLink connectivity, memory architecture, packaging, manufacturing and rack-scale systems.

Customers can bring their XPU designs to MediaTek and configure connectivity, memory, packaging, performance and power characteristics according to their workload and infrastructure requirements.

MediaTek is also part of the NVIDIA NVLink Fusion ecosystem, which supports connections between custom XPUs and NVIDIA AI infrastructure for hyperscalers, cloud service providers and AI model developers.

Through NVLink Fusion, partners can use NVIDIA scale-up and scale-out technologies and the NVIDIA MGX rack-scale architecture as part of their custom AI infrastructure development.

Supporting local AI computing

NVIDIA and MediaTek are continuing their collaboration on local AI computing for generative and agentic AI applications, combining NVIDIA accelerated computing with MediaTek’s system-on-chip design capabilities.

MediaTek collaborated with NVIDIA on the GB10 Grace Blackwell Superchip used in NVIDIA DGX Spark. The system combines an NVIDIA Blackwell GPU and Grace CPU through NVLink-C2C for local AI computing. The companies have also extended their collaboration to NVIDIA RTX Spark platforms for consumer PCs.

Developing AI-enabled vehicle platforms

NVIDIA and MediaTek are also collaborating on multiple generations of platforms for AI-enabled, software-defined vehicles. The work combines MediaTek’s automotive SoC design capabilities with NVIDIA’s accelerated computing, AI, graphics and software technologies.

MediaTek Dimensity Auto platforms integrate NVIDIA technologies for AI and NVIDIA RTX graphics in vehicle cockpits and can operate alongside NVIDIA DRIVE AGX. The companies plan to continue developing the architecture for future vehicle platforms with increased AI capabilities.

For more information, visit nvidia.com.

The post NVIDIA and MediaTek expand collaboration across AI and autos appeared first on Engineering.com.