NVIDIA Invests $2 Billion in Marvell to Deepen AI Data-Center, Silicon Photonics Ties

Cloud providers and artificial-intelligence developers are increasingly designing custom accelerators, challenging Nvidia to expand beyond selling GPUs into platforms, interconnects and chip partnerships. MediaTek’s system-on-chip design and mass-production expertise gives Nvidia a route into the growing ASIC market, while broadening an existing relationship spanning AI data centers, personal computers, smart vehicles and high-performance computing systems.

As of Sept. 2, 2026, Nvidia has announced a $3.5 billion investment through the purchase of MediaTek convertible bonds, tightening financial and technical ties between the companies. MediaTek will adopt Nvidia’s NVLink Fusion platform to help hyperscalers and AI model developers design, integrate and manufacture custom accelerators. The partners also plan to advance in-house AI chips and chip-to-chip connectivity, positioning the alliance to serve Big Tech’s expanding data-center buildout.

The rapid growth of generative AI models and cluster sizes is exposing bandwidth, latency and power-consumption constraints in copper-based data transmission between GPUs. Silicon photonics transmits signals using light, while co-packaged optics (CPO) places optical engines closer to switch chips, improving bandwidth and energy efficiency. NVIDIA’s push could also create opportunities for Taiwanese suppliers involved in technologies such as TSMC’s COUPE platform and ASE Technology’s advanced packaging.

On March 2, 2026, NVIDIA announced multiyear strategic agreements with Lumentum and Coherent, investing $2 billion in each company for a combined $4 billion. The agreements cover research and development, U.S. manufacturing capacity and operations, and include purchase commitments and rights to advanced optical-product capacity. After NVIDIA invested another $2 billion in Marvell on March 31, its related optical and silicon-photonics investments rose to $6 billion.