NVIDIA invests $3.5 billion in MediaTek convertible bonds
NVIDIA invested $3.5 billion in MediaTek convertible bonds as MediaTek adopts NVLink Fusion as the foundation for custom AI accelerators.
NVIDIA invested $3.5 billion in convertible bonds issued by MediaTek as part of an expanded partnership under which MediaTek will adopt NVIDIA’s NVLink Fusion platform as a design foundation for custom AI accelerators. The companies announced the agreement on August 31, 2026.
The agreement extends NVIDIA’s rack-scale interconnect to accelerators NVIDIA does not design itself. MediaTek will be able to connect custom XPUs — accelerators built to a single customer’s specification rather than sold as a general-purpose part — to NVIDIA NVLink-connected rack-scale systems.
NVIDIA said NVLink Fusion includes a chiplet for scale-up fabric connectivity, NVLink-C2C for processor connectivity and NVHBM for customized memory capabilities. The companies said the prebuilt and prevalidated foundation supplies connectivity, memory architecture, packaging, manufacturing and rack-scale technologies around a customer’s differentiated compute. They said it could reduce development complexity and speed time to market, but published no data demonstrating those expected benefits.
“Together, we’re building platforms that bring NVIDIA accelerated computing to new markets and give customers the freedom to create differentiated AI systems at enormous scale,” NVIDIA founder and CEO Jensen Huang said.
The expanded collaboration covers cloud AI infrastructure, local AI computing and automotive. The companies said it continues multigenerational work on NVIDIA RTX Spark and DGX Spark PC chips and on platforms for AI-powered, software-defined vehicles. For context on a separate financing mechanism, DataPhoenix previously covered NVIDIA’s revenue-sharing financing model for AI cloud partners.
The announcement did not disclose the bonds’ coupon, maturity, conversion price or ratio, issuance date, governing conditions, or the ownership position NVIDIA would hold on conversion. It also did not set a deployment schedule, name MediaTek custom-accelerator customers, quantify expected performance gains, or confirm customer adoption of NVLink Fusion-based designs. NVIDIA said statements about future plans and expected benefits are subject to risks and uncertainties that could cause actual results to differ materially.
More news

AWS releases six open-source Hugging Face deployment skills for SageMaker

Google Research releases MilleMiglia logistics benchmark generator

AWS launches AgentCore Runtime V2 with elastic memory and snapshot starts
