NextFin News - Nvidia CEO Jensen Huang has unveiled a state-of-the-art campus in Taiwan, marking a dramatic escalation in the silicon giant's efforts to anchor its next-generation artificial intelligence hardware ecosystem on the island. The move comes just a week after rival Advanced Micro Devices, led by CEO Lisa Su, announced its own plans to invest more than $10 billion in Taiwan's AI sector. According to details shared by AI researcher and industry commentator Rohan Paul on X, Nvidia is planning an annual investment of around $150 billion in Taiwan to support this expansion. While the announcement cements Taiwan's position as the undisputed epicenter of the global AI supply chain, the staggering scale of the reported investment has immediately triggered intense debate among financial analysts regarding its currency denomination and operational feasibility.
Paul, who has built a reputation for his highly optimistic and bullish stance on AI infrastructure and hardware expansion, presented the $150 billion figure as a testament to Nvidia's aggressive scaling strategy. However, this projection currently stands as an isolated data point and lacks cross-validation from official corporate filings or sell-side consensus. Financial analysts have quickly pointed out that if denominated in U.S. dollars, a $150 billion annual commitment would exceed Nvidia’s entire projected capital expenditure and even its annual revenue, making it highly improbable. Instead, local market observers suggest the figure likely refers to New Taiwan Dollars (NTD). An annual investment of 150 billion NTD translates to roughly $4.6 billion USD—a figure that is highly realistic, yet still represents a monumental capital commitment that would comfortably outpace AMD's recently announced plans.
The timing of Huang's presentation is far from coincidental. Just days earlier, AMD committed over $10 billion to establish its own R&D facilities and joint engineering centers in Taiwan. The rapid-fire announcements highlight how the battle between the two chip designers has moved beyond architecture design and into the physical securing of supply chain capacity. Taiwan, home to Taiwan Semiconductor Manufacturing Company (TSMC), controls over 90% of the world's advanced logic chip manufacturing and virtually all of the advanced packaging capacity, such as Chip-on-Wafer-on-Substrate (CoWoS), which is critical for both Nvidia's Blackwell and AMD's Instinct MI300 series. By establishing dedicated physical campuses, both companies are attempting to build deeper, localized engineering alliances with TSMC and key assembly partners like Foxconn and Quanta Computer.
Yet, this aggressive concentration of capital in a single geographic region carries significant operational and geopolitical risks. Critics of the heavy Taiwan-centric strategy argue that both Nvidia and AMD are doubling down on a single point of failure. Supply chain bottlenecks, local power grid stability, and water shortages have frequently challenged Taiwan's industrial parks. Furthermore, the sheer volume of talent required to staff these new campuses is stretching the island's engineering pool to its limits. While Nvidia's brand and financial clout give it a distinct advantage in recruiting, the talent war with AMD, local design houses, and TSMC itself could drive up operational costs and dilute engineering efficiency.
Broader market indicators suggest some institutional investors remain cautious about whether these massive capital commitments will yield immediate returns. While demand for AI chips remains robust, there are growing concerns about the monetization timeline for enterprise AI software. If hyperscalers begin to slow their capital expenditures on data centers, Nvidia's massive localized investments could lead to temporary overcapacity. For now, however, the momentum remains firmly behind physical expansion. Huang's showcase of the new campus serves as a powerful signal to both competitors and partners that Nvidia has no intention of relinquishing its dominant market share, choosing instead to build physical and engineering moats that will be incredibly difficult for rivals to breach.
Explore more exclusive insights at nextfin.ai.
