NVIDIA Expands Taiwan R&D Beyond GPUs, Designing Seven Types of Chips

NVIDIA is significantly expanding its research and development footprint in Taiwan, moving beyond its traditional focus on graphics processing units (GPUs) to design and develop seven different types of semiconductors, a shift that underscores the island's growing role in the company's global AI computing strategy.

NVIDIA Chief Executive Officer Jensen Huang (黃仁勳) confirmed the expansion on Thursday while addressing more than 2,000 employees at the company's annual year-end gathering in Taipei. He said Taiwan has evolved from a single-product R&D base into a multi-chip development hub supporting NVIDIA's increasingly complex computing ecosystem.

“People may still think of NVIDIA as a GPU company,” Huang said. “But our GPUs today function more like supercomputers. That requires a much broader system architecture and a much larger engineering effort.” （Related： Ex‑TSMC R&D Chief: Without EUV, China’s Chip Industry Is “Fighting With One Leg Broken” ｜ Latest ）

From GPUs to a Full-Stack Chip Portfolio

According to Huang, NVIDIA now designs and develops seven categories of chips in Taiwan, including networking chips, switch chips, smart data processors for storage and networking applications, and central processing units (CPUs), in addition to its core GPU products.

The expansion reflects NVIDIA's transition from a graphics-focused chipmaker into what Huang describes as a global AI computing company, as demand accelerates for integrated systems capable of handling large-scale artificial intelligence workloads.

As system complexity increases, Huang said NVIDIA requires a growing pool of engineering talent in areas such as chip design, system architecture, software development, and supply chain coordination—capabilities that Taiwan's semiconductor ecosystem is well positioned to provide.

MediaTek Partnership Targets AI PC Market

Huang also confirmed for the first time that NVIDIA is working closely with MediaTek on a new system-on-chip project, widely known in the market as the N1 processor. He described the chip as a high-performance, low-power design aimed at next-generation personal computers with advanced AI computing capabilities.

The collaboration signals a joint push by the two companies into the emerging AI PC market, where chipmakers are racing to redefine computing performance beyond traditional CPUs and GPUs. （Related： Ex‑TSMC R&D Chief: Without EUV, China’s Chip Industry Is “Fighting With One Leg Broken” ｜ Latest ）

Deepening Ties with Taiwan's Semiconductor Ecosystem

During his visit, Huang emphasized the strategic importance of Taiwan to NVIDIA's long-term development. He revealed that his first formal meeting after arriving in Taiwan was a dinner with Taiwan Semiconductor Manufacturing Company founder Morris Chang (張忠謀).

“I'm very happy to see Morris,” Huang said. “I feel deeply grateful to him.”

Despite the weekend timing, Huang said his schedule was filled with meetings and dinners with Taiwan-based supply chain partners, highlighting NVIDIA's close integration with the island's semiconductor industry.

Looking Beyond AI

Beyond current AI applications, Huang also pointed to NVIDIA’s interest in quantum computing. He noted that the company recently released a blueprint for the U.S. National Quantum Initiative, outlining plans to combine GPUs with quantum processing units (QPUs) to simulate complex biological and physical systems.

Huang added that Taiwan’s scientific research capacity positions it as a potential partner in exploring the next frontier of computing beyond artificial intelligence.

（Related： Ex‑TSMC R&D Chief: Without EUV, China’s Chip Industry Is “Fighting With One Leg Broken” ｜ Latest ）

​