Photo: ed br / Pexels
The two chip giants are building a shared AI ecosystem spanning cars, PCs, and data centers, with a multibillion-dollar financial commitment to match.
Nvidia and MediaTek announced a partnership on June 1, 2026, extending across AI-powered laptops, automotive cockpits, and data-center infrastructure. Reports have placed Nvidia’s financial commitment at somewhere between $3.5B and $4B in MediaTek convertible bonds, though those figures remain unconfirmed by either company.
What they’re building together #
The most consumer-visible product to emerge so far is the NVIDIA RTX Spark, a platform designed specifically for Windows 11 AI PCs, pairing Nvidia’s GPU capabilities with MediaTek’s expertise in system-on-chip design. First laptops featuring RTX Spark are expected to hit shelves in Fall 2026.
On the automotive side, the collaboration integrates MediaTek’s 3nm Dimensity AX C-X1 platform into AI-enabled cockpit systems.
MediaTek co-designed the NVIDIA GB10 Grace Blackwell Superchip at the heart of Project DIGITS, Nvidia’s personal AI supercomputer.
MediaTek has also stood up its own on-premises AI factory running on Nvidia’s DGX SuperPOD systems. The factory processes roughly 60 billion inference tokens every month.
The money behind the move #
MediaTek’s board approved $5B in discretionary financing as of July 31, 2026, to fund the company’s push into data-center AI chips. The company’s AI chip business is projected to generate over $2B in revenue for 2026, with the first custom ASIC rolling off production lines in Q4 2026 and a second custom ASIC slated for 2028.
Why the timing matters #
The broad addressable market for custom AI chips is projected to reach around $80B by 2027. MediaTek, with Nvidia’s backing, is positioning itself as a partner for companies that want custom AI silicon, supported by Nvidia’s software stack including its CUDA libraries and NeMo platform for LLM development.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our