Nvidia's AI Dominance Continues Unchecked as Trump Softens on Chipmaker

Nvidia, a gaming pure-play turned AI kingmaker, has reached a $4 trillion milestone. The company's success is driven by its robust H100 and H200 Tensor Core GPUs powering the global AI boom. Nvidia's quarterly top-line numbers have tripled, and the stock has surged with a consensus of 38 Wall Street analysts predicting a $182.49 price target over the next year, with some targets as high as $250. The White House's support, including rolling back export restrictions and loosening permitting rules, is seen as a critical tailwind for Nvidia's growth.

Nvidia, the gaming pure-play turned AI kingmaker, has reached a $4 trillion milestone, driven by its robust H100 and H200 Tensor Core GPUs. The company's quarterly top-line numbers have tripled, and the stock has surged, with a consensus of 38 Wall Street analysts predicting a $182.49 price target over the next year, with some targets as high as $250. The White House's support, including rolling back export restrictions and loosening permitting rules, is seen as a critical tailwind for Nvidia's growth.

Despite the challenges posed by U.S. export restrictions, Nvidia's advanced GPUs remain in high demand. In China, a thriving repair industry has emerged, with around a dozen boutique companies offering repair services for Nvidia's AI chipsets [1]. This industry supports the view that there has been significant smuggling of Nvidia chipsets into China, with the government and military making purchases of the banned AI chips [1]. The repair demand is so high that some companies can repair up to 500 Nvidia AI chips per month [1].

The Financial Times recently reported that at least $1 billion worth of Nvidia GPUs were shipped to China during a three-month period when export controls were being tightened [2]. The investigation found that Chinese distributors sold multiple models of Nvidia AI chips, including the B200, H100, and H200, in potentially vast numbers. The chips were allegedly sold in racks and packaged with the components and software needed to hook them directly into an existing data center [2].

Nvidia has only just been allowed to recommence sales of its H20 AI chipset in China, which has been specifically developed to comply with U.S. restrictions. However, switching over to H20 chipsets is not necessarily a simple or good option for Chinese entities due to price and performance considerations [1]. Some firms involved in the training of large language models would likely prefer H100 chipsets, which are better suited to that task [1].

The White House's recent easing of restrictions on Nvidia chip exports may help reduce the need for black markets. Nvidia has filed applications to sell the H20 GPU in China once more, with assurance from the U.S. government that the licenses will be granted [2]. This could make the need for some black markets a thing of the past.

Nvidia's success is not just in China. The company's industry-leading GPUs provide the computational horsepower that underpins artificial intelligence (AI), and experts predict there's a long runway ahead for the adoption of AI [3]. The biggest purveyors of the technology, including Amazon, Microsoft, Alphabet, and Meta Platforms, plan to spend hundreds of billions of dollars on capital expenditures (capex) in 2025 to support their AI ambitions [3]. Nvidia is likely to be on the receiving end of a significant portion of that spending.

References:
[1] https://finance.yahoo.com/news/china-repair-demand-banned-nvidia-210306490.html
[2] https://www.inkl.com/news/financial-times-report-suggests-chip-smugglers-shifted-an-alleged-1-billion-worth-of-nvidia-s-ai-chips-to-china-over-the-course-of-3-months-while-stricter-us-export-controls-were-in-effect
[3] https://finance.yahoo.com/news/prediction-supercharged-growth-stock-soar-070200102.html