TLDR: Nvidia and TSMC have officially begun producing advanced Blackwell AI wafers in the United States, primarily at TSMC’s Arizona facilities, with an anticipated investment of up to $500 billion in AI infrastructure over the next four years. This landmark development aims to bolster American technological sovereignty, enhance the domestic AI supply chain, and is projected to create hundreds of thousands of jobs, significantly reshaping the global tech landscape.
In a pivotal moment for the global technology industry and a substantial leap towards fortifying American technological sovereignty, Nvidia and Taiwan Semiconductor Manufacturing Company (TSMC) have officially commenced the production of advanced AI chips within the United States. The unveiling of the first US-made Blackwell wafer in October 2025 signifies a strategic realignment in the semiconductor supply chain and a robust commitment to domestic manufacturing for the burgeoning artificial intelligence sector.
Nvidia, in collaboration with TSMC, is fabricating these cutting-edge Blackwell wafers at TSMC’s advanced Phoenix facilities in Arizona. This initiative is part of Nvidia’s broader strategy to localize its AI supercomputer production, with plans to invest up to $500 billion in AI infrastructure in the U.S. over the next four years. Jensen Huang, CEO of Nvidia, stated, ‘The engines of the world’s AI infrastructure are being built in the United States for the first time.’ He further added, ‘Adding American manufacturing helps us better meet the incredible and growing demand for AI chips and supercomputers, strengthens our supply chain and boosts our resiliency.’
The Blackwell wafer itself represents a significant technological advancement. It is built on a custom TSMC 4NP process and features an astonishing 208 billion transistors, more than 2.5 times the 80 billion found in its Hopper predecessor. This dual-die configuration, where two reticle-limited dies are seamlessly connected by a blazing 10 TB/s NV-High Bandwidth Interface (NV-HBI), allows them to function as a single, cohesive GPU, delivering unparalleled computational density and efficiency. A standout innovation is the incorporation of FP4 (4-bit floating point) precision, which effectively doubles the performance and memory support for next-generation models while rigorously maintaining high accuracy in AI computations.
Beyond wafer production, Nvidia is expanding its domestic manufacturing ecosystem. The company is partnering with Amkor and Silicon Precision Industries (SPIL) for packaging and testing operations in Arizona, with Amkor having announced a $2 billion packaging facility in the state. Additionally, Nvidia is collaborating with contract electronics manufacturers Foxconn in Houston and Wistron in Dallas to establish two supercomputer manufacturing plants in Texas. Production at these sites is expected to ramp up over the next 12 to 15 months.
Also Read:
- Nvidia’s DGX Spark: Jensen Huang Personally Delivers Compact AI Supercomputers to Elon Musk and Sam Altman
- Nvidia CEO Jensen Huang Highlights AI’s Unprecedented Complexity, Predicts Elon Musk’s Gigawatt AI Compute Leadership
This monumental development is poised to profoundly reshape the global AI ecosystem. It primarily benefits American AI and technology innovation companies, particularly those at the forefront of large language models (LLMs) and generative AI. The move is expected to create hundreds of thousands of jobs and drive trillions of dollars in economic security over the coming decades, aligning with the objectives of the CHIPS Act to reduce dependence on overseas facilities and enhance resilience against geopolitical tensions. While Nvidia’s B200 GPUs are already shipping, the company anticipates the Blackwell Ultra (B300-series) in the second half of 2025, promising an approximate 1.5x speed increase over the base Blackwell model.
