The two companies announced in a joint statement on Monday that Nvidia will purchase bonds convertible into MediaTek stock.
This investment expands the cooperation between two chip design companies. According to the agreement, MediaTek will use Nvidia's NVLink Fusion and newly released NVHBM technology as part of its technology suite to help facilitate smoother communication between data center components. This smaller company is attempting to challenge Broadcom and Marvel Technologies by helping data center owners create their own components.
For Nvidia, this new agreement is its latest move to ensure that the future of artificial intelligence is built around its hardware platform and standards. The company has been building data center interconnect systems, such as NVLink Fusion, and collaborating with companies like MediaTek to ensure its core position throughout the entire hardware chain, not just limited to its leading AI accelerators.
Previously, Amazon also announced a similar agreement, agreeing to deploy an additional 2 million Nvidia components. More importantly, Amazon has also promised to use Nvidia's connectivity technology in its self-developed chips.
Nvidia CEO Huang Renxun, born in Taiwan, is strengthening ties with a company committed to reducing its dependence on the smartphone market. Earlier this year, MediaTek partnered with Google, a subsidiary of Alphabet, and gained increasing recognition as a design partner for a large technology company seeking to build its own AI chips. This development momentum has led to a nearly two-fold increase in the market value of this chip manufacturer in recent months.
The agreements with MediaTek and Amazon give Nvidia the opportunity to maintain its core position in the construction of artificial intelligence data centers, even as the company deploys chips aimed at replacing its main products. This transaction once again demonstrates Nvidia's investment as an industry leader in customers, competitors, and other AI related enterprises. Despite Nvidia's strong denial that it is creating false demand, there are still concerns among investors about these seemingly cyclical investments.
The management of the company stated that the huge investment is just another way to accelerate the popularization of its technology and will drive the development of the artificial intelligence industry. Nvidia remains a major beneficiary of significant investments by some of the world's largest companies in artificial intelligence data centers. Its acceleration chip derived from graphics processors performs well in creating and running artificial intelligence tools and services. With this advantage, Nvidia is constantly developing other technologies including complete computers and networks.
Nvidia's leading position has attracted numerous potential competitors and prompted its largest customers, including Amazon, Google, and Microsoft, to increase their internal chip research and development efforts. At present, Nvidia's dominant position remains solid: the company predicts a significant increase of 70% in sales next year.
MediaTek has partnered with Nvidia to help this American chip manufacturer enter the laptop market. This Taiwanese company is assisting Nvidia in developing RTX Spark products for this market.
MediaTek also stated that it plans to actively recruit talents to better seize the business opportunities brought by the artificial intelligence boom. The company expects revenue from artificial intelligence chips to reach approximately $2 billion this year and has set a goal of occupying 15% of the $80 billion data center market next year.
NVIDIA and MediaTek deepen long-term partnership to jointly build an artificial intelligence platform from edge to cloud computing
NVIDIA and MediaTek announced today the deepening of their long-term cooperation to jointly build the next generation of artificial intelligence computing platforms, covering areas such as artificial intelligence infrastructure, local artificial intelligence computing, and automotive applications.
As part of this collaboration expansion, MediaTek will adopt NVIDIA NVLink Fusion ™ The platform provides a pre validation path for ultra large scale data centers, cloud service providers, and cutting-edge model developers to develop customized XPUs and integrate them into NVIDIA NVLink ™ Connected rack mounted artificial intelligence factories.
Both parties will combine NVIDIA's accelerated computing, artificial intelligence, graphics, and software platforms with MediaTek's leading advantages in custom chips, high-performance computing, energy-efficient System on Chip (SoC) design, advanced packaging, interconnect, and connectivity technologies.
NVIDIA also invested $3.5 billion to purchase convertible bonds issued by MediaTek.
NVIDIA and MediaTek are collaborating in three major areas:
Artificial Intelligence Infrastructure: MediaTek will collaborate with NVIDIA's NVLink Fusion ecosystem to help customers develop customized AI infrastructure for integration with NVIDIA rack level systems and AI factories.
Local Artificial Intelligence Computing: Both parties will continue to collaborate on the development of multiple generations of NVIDIA RTX Spark ™ DGX Spark ™ PC chips integrate NVIDIA GPUs with MediaTek SoCs to power consumer PCs, artificial intelligence developer supercomputers, and enterprise workstations.
In the automotive field, MediaTek and NVIDIA will continue to develop software defined automotive platforms driven by artificial intelligence to welcome the arrival of the era of physical artificial intelligence.
Artificial intelligence is changing every computing platform - from the world's largest artificial intelligence factory to personal computers and cars, "said Huang Renxun, founder and CEO of NVIDIA. MediaTek is one of the world's leading semiconductor companies with extraordinary expertise in system on chip design, connectivity, superior performance, and energy efficiency. We will work together to build a platform that brings NVIDIA's accelerated computing to new markets and allows customers to freely create differentiated AI systems on a large scale. ”
MediaTek and NVIDIA share a common vision of popularizing advanced AI computing across the entire technology field, "said Cai Lixing, Vice Chairman and CEO of MediaTek. NVIDIA's investment strengthens our collaboration in the automotive industry in the era of cloud AI infrastructure, local AI computing, and physical AI. By combining NVIDIA's leading position in accelerating computing and AI software ecosystems with MediaTek's expertise in a diverse AI technology portfolio from edge to cloud, and our leading position in custom chips, we can accelerate customer innovation. ”
Building customized AI infrastructure using NVLink Fusion platform
MediaTek will provide the NVLink Fusion platform as the design foundation for custom AI accelerators developed by customers, enabling their platform to develop in sync with future NVIDIA architectures.
The NVLink Fusion platform provides a pre built, pre certified, and system pre validated infrastructure for multi chip XPU development, accelerating the development process from chip and advanced packaging to rack level systems.
The NVLink Fusion platform integrates key technologies for customized XPUs, including:
The NVIDIA NVLink Fusion chipset utilizes NVIDIA photon or electrical interconnect technology to connect XPUs to the NVIDIA NVLink expansion architecture.
NVIDIA NVLink-C2C， Provide high bandwidth, energy-efficient connections between XPU, NVIDIA Rosa CPU, and other compatible processors.
NVIDIA NVHBM， Integrate customized memory functionality to increase bandwidth and energy efficiency while allocating more chip area for computing.
Building custom accelerators is just the beginning of deploying custom XPUs in rack level AI factories. Integrating multi chip architecture, advanced packaging, high-speed SerDes, HBM, I/O, and scalable networks into a manufacturable and mass-produced system requires extensive engineering design, certification, and supply chain support from chip to rack.
Customers do not need to design and certify every component of custom XPU from scratch, but can focus their resources on defining the differentiated computing capabilities of their platform, while relying on NVLink connectivity, memory architecture, packaging, manufacturing, and rack level technologies provided by NVIDIA and MediaTek to achieve production deployment.
Customers can submit their XPU designs to MediaTek and customize connectivity, memory, packaging, performance, and power characteristics based on their workload and infrastructure requirements.
In addition, MediaTek is also part of the broader NVIDIA NVLink Fusion ecosystem, which enables hyperscale data centers, cloud service providers, and cutting-edge model developers to seamlessly connect custom XPUs to NVIDIA's AI infrastructure.
With NVLink Fusion, partners can leverage NVIDIA's mature vertical and horizontal scaling technology stack and ecosystem, as well as NVIDIA MGX ™ Rack level architecture reduces development complexity, improves performance, and accelerates the time to market for semi custom AI factories.
Introducing supercomputers into local AI computing
NVIDIA and MediaTek are working together to advance local AI computing to embrace the era of generative and intelligent AI. Both parties will combine NVIDIA's leading position in accelerated computing with MediaTek's expertise in high-performance, energy-efficient system level chip design.
MediaTek has collaborated with NVIDIA to develop the GB10 Grace Blackwell supercomputer, which powers the NVIDIA DGX Spark. The NVIDIA DGX Spark integrates NVIDIA Blackwell GPU and Grace CPU, and is connected through NVLink-C2C to bring powerful AI capabilities to edge systems. The two companies have expanded their collaboration to NVIDIA RTX Spark, aiming to power the next generation of consumer PCs for the AI era.
Advancing the development of AI defined cars in the era of physical AI.
In addition, NVIDIA and MediaTek are collaborating across multiple generations to advance the development of AI driven software defined cars. Both parties will leverage MediaTek's leading position in automotive system level chip design and NVIDIA's expertise in accelerating computing AI、 Combining advantages in graphics and software.
The MediaTek Dimensity automatic platform integrates NVIDIA technology, providing advanced AI and NVIDIA RTX for intelligent car cockpits ™ Graphics processing capability and compatible with NVIDIA DRIVE AGX ™ Collaborative work. The two companies will continue to build scalable architectures for future generations of products based on this foundation to support increasingly intelligent AI vehicles.
Subscribe And Save Stay in touch.
