Every nation and business aspires to expand and generate economic opportunities, but doing so requires almost infinite knowledge. This week, NVIDIA, in collaboration with its ecosystem partners, is highlighting its efforts to develop reasoning, AI models, and compute infrastructure to produce intelligence in AI factories that will propel the next stage of growth in the United States and beyond.
NVIDIA declared that it will begin producing NVIDIA AI supercomputers in the United States. The corporation and its partners intend to build up to half a trillion dollars’ worth of AI infrastructure in the United States over the course of the next four years.
NVIDIA to Make First US-Made AI Supercomputers
NVIDIA and its manufacturing partners are designing and building US factories to make all AI supercomputers for the first time.
In order to create and test AI supercomputers in Texas and NVIDIA Blackwell chips in Arizona, the business has put more than a million square feet of manufacturing facilities into service with leading manufacturing partners.
At TSMC’s chip plants in Phoenix, Arizona, NVIDIA Blackwell chips production has started. NVIDIA is building supercomputer production facilities in Texas, joining Foxconn in Houston and Wistron in Dallas. Both plants are expected to enhance their mass output in the next 12 to 15 months.
Supercomputers and AI chips have a complex supply chain that necessitates the use of the most advanced technology for production, testing, packaging, and assembly. NVIDIA is working with Amkor and SPIL in Arizona on packaging and testing projects.
Over the next four years, NVIDIA plans to invest up to half a trillion dollars in AI infrastructure in the United States through partnerships with TSMC, Foxconn, Wistron, Amkor, and SPIL. These pioneers in the sector are growing their businesses, expanding globally, fortifying their supply chains, and fortifying their partnership with NVIDIA.
NVIDIA AI supercomputers power a new type of data center built especially to manage artificial intelligence, and AI factories form the backbone of a new AI economy. Tens of “gigawatt AI factories” are expected to be built in the next years. Production of NVIDIA Blackwell chips and supercomputers for American AI manufacturers is expected to generate trillions of dollars in economic security and hundreds of thousands of jobs over the next few decades.
Jensen Huang, the founder and CEO of NVIDIA, claims that “the engines of the world’s AI infrastructure are being built in the United States for the first time.” “It can better meet the amazing and expanding demand for AI chips and supercomputers by strengthening the supply chain and increasing resilience by incorporating American manufacturing.”
Using its state-of-the-art AI, robotics, and digital twin technologies such as NVIDIA Omniverse to create digital twins of factories and NVIDIA Isaac GR00T to build robots to automate manufacturing the company will build and operate the facilities.
Over the next several decades, it is anticipated that the construction of NVIDIA AI supercomputers in the United States for American AI factories would generate employment possibilities for hundreds of thousands of people and generate trillions of dollars in growth. TSMC fabs in Arizona are already producing some of the NVIDIA Blackwell computing engines that power those AI supercomputers.
Today, NVIDIA announced that CoreWeave now offers NVIDIA Blackwell GB200 NVL72 rack-scale systems enabling users to run applications at scale and train next-generation AI models. To train and implement the next generation of AI, CoreWeave now has hundreds of NVIDIA Grace Blackwell processors accessible.
NVIDIA leads the way in AI software development to provide more intelligent and effective models in addition to hardware advancements.
The NVIDIA Llama Nemotron Ultra model, which represents the most recent of those developments, was named the most accurate open-source reasoning model for complicated coding and scientific problems by Artificial Analysis. Additionally, it is currently regarded as one of the best reasoning models available.
It is all based on the engineering achievements of NVIDIA. After competing against 2,200 teams to solve challenging mathematical reasoning problems which are essential to furthering scientific discovery, disciplines, and domains a team of NVIDIA engineers took first place in the AI Mathematical Olympiad. The Llama Nemotron Ultra model was trained using the identical post-training methods and publicly available datasets as NVIDIA’s winning entry in the math reasoning contest.
There is an almost infinite demand for intelligence in the world, and NVIDIA’s AI platform is simultaneously helping to supply that want everywhere.