Urgent.News

What's breaking now, across thousands of outlets.

AI

Nvidia is investing $3.5 billion in Taiwan's MediaTek to deepen AI chip partnership

Nvidia will buy convertible bonds in the Taiwanese chipmaker as the two companies broaden collaboration across data centers, PCs, and automotive

Nvidia is investing $3.5 billion in Taiwan's MediaTek to deepen AI chip partnership

Nvidia has invested $3.5 billion in MediaTek through convertible bonds as part of an expanded partnership. A convertible bond allows the holder to convert it into shares later on. The companies announced this expanded partnership today. Nvidia regularly purchases stakes in partners as part of technical collaborations. Earlier this year, Nvidia bought $2 billion worth of shares in Marvell Technologies, a competitor of MediaTek.

MediaTek is known for supplying smartphone and connected device processors, and it also has a custom chip design business specializing in artificial intelligence accelerators. Nvidia's AI unit, which is the first focus of the expanded partnership, offers a catalog of on-chip networking technologies for AI accelerator design. MediaTek will offer Nvidia's NVLink Fusion networking technology as an interconnect option for AI racks, which contain not only graphics processing units but also central processing units.

NVLink Fusion enables Nvidia's GPUs and CPUs to exchange data with third-party CPUs and GPUs at speeds 10 times faster than standard PCIe 6.0 implementations. The company's partnership with MediaTek also prioritizes the personal computing market and the auto sector. In 2025, Nvidia launched the DGX Spark workstation, powered by a MediaTek-designed system-on-chip, and the RTX Spark, a laptop and desktop processor.

MediaTek helped develop the RTX Spark's onboard CPUs and certain other components, such as the memory controller. The companies plan to collaborate on multiple future generations of the RTX Spark and DGX Spark. Additionally, MediaTek sells Dimensity Auto systems-on-chips for in-vehicle displays, which include an Nvidia graphics rendering chiplet, and the two companies plan to continue working together on software-defined vehicle chips.

Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at qz.com →

More in AI

3 Costly Mistakes I Made With the OpenAI API (So You Don't Have To)

You get the first bill from OpenAI and it's 10x what you expected. We've all been there. It's not about the cost of a single call. It's about the thousands of calls you didn't know you were making.

  • Ignored Temperature and Max Tokens, leading to unexpected high costs
  • Lacked user-specific cost tracking, hindering financial oversight
  • Neglected model version differences, incurring unnecessary expenses

MUSTER: Building Safer Enterprise AI Agents That Ask Less and Retry Less

AI agents are getting good enough to do real work inside companies. They can read documents, understand images, call tools, make decisions, and trigger actions.

  • MUSTER system limits agent's data access and verifies action outcomes
  • Each data source controls its evidence, central system has no blanket access
  • MUSTER minimizes data collection when uncertainty exists, avoiding unnecessary data gathering

More from Monday 31 August →