Nvidia Pours $3.5B Into MediaTek for AI Chip Tie-Up — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Nvidia is investing $3.5 billion into MediaTek, which will adopt Nvidia's NVLink Fusion technology to design custom AI chips for hyperscalers that plug directly into Nvidia-based data centers.
- MediaTek expects its custom data-center ASIC business to generate $2 billion in revenue in 2026, per a company statement from June.
- Nvidia struck a parallel partnership with AWS last week—without a direct equity investment—committing 2 million additional GPUs and integrating the same NVLink Fusion stack.
- The strategy responds to a chip-buildout push by Amazon, Google, Microsoft, OpenAI, and Anthropic, all of which are designing in-house silicon to reduce dependence on Nvidia GPUs.
- The deal goes beyond data centers: the companies will continue collaborating on DGX Spark (small developer desktops), RTX Spark (consumer AI PCs), and AI-powered software-defined vehicles using Nvidia Drive AGX.
- Nvidia's Dion Harris, senior director of HPC and AI hyperscaler infrastructure solutions, positioned the company as "an AI infrastructure company" that "expanded beyond pure computing chips years ago."
- Jensen Huang, Nvidia's founder and CEO, said the partnership targets "every computing platform—from the world's largest AI factories to the PC and the car."
Why it matters: Nvidia is essentially monetizing the connective tissue—NVLink Fusion—while ceding the chip itself, a tradeoff that preserves its data-center dominance as $2 billion-revenue custom-silicon programs at MediaTek scale up. Investors in MediaTek gain a premium channel into hyperscaler buyers; Nvidia's GPU competitors gain less leverage than they would have without the Fusion standard binding custom silicon back to Nvidia racks.
Ask SkimNews
