On Wednesday, semiconductor chip giant Nvidia reported strong earnings and revenues for its fourth quarter ended Jan. 31. Revenues for the quarter were up 78% to $39.33 billion from the year ago quarter. Net income growth was also impressive, $22.09 billion or 89 cents a share, against $12.29 billion or 49 cents a share, year on year. That is an earnings per share growth rate exceeding 70%. The company raised guidance for the current quarter.
Jensen Huang, Nvidia’s CEO, said demand for Blackwell, the company’s newest semiconductor platform, is “amazing.” Huang added, “We’ve successfully ramped up the massive-scale production of Blackwell AI supercomputers, achieving billions of dollars in sales in its first quarter.”
The future remains bright for what is arguably the world’s most important company. Nvidia’s advanced graphics processing chips, GPUs are at the center of the AI revolution. Companies are rapidly embracing artificial intelligence in order to improve productivity. National productivity is also accelerating. Other nations, including Japan and the larger countries of Western Europe, are racing to catch up in the race to embed AI in their economies. The race for AI supremacy is intense.
China is also aggressively investing in AI. In 2024, Chinese businesses, all of which are ultimately servants of the ruling Communist Party, purchased over one million Nvidia H20 semiconductors. This generated about $12 billion in revenues for Nvidia. The H20 is designed specifically to meet U.S. restrictions on exports of accelerated computing semiconductors to China.
The majority of Nvidia’s most advanced GPUs are being purchased by the so-called American hyperscalers, Alphabet-Google, Amazon, Microsoft, and Meta. The first three named companies will invest about $80 billion each in AI data centers this year. This week, Meta announced that it would invest $200 billion in an AI data center campus. Translation? Talk of a slowdown in AI spending by the hyperscalers is nonsense. Competition is intense among these four companies. Rationally, each company will invest as much as possible in order not to be left behind.
At the moment, Nvidia is supply constrained by fabrication capacity limits at its sole provider of advanced AI chips, Taiwan Semiconductor. As Taiwan Semiconductor expands capacity, nations will become an additional source of demand for Nvidia’s most advanced semiconductor chip platform, the Blackwell platform. Each Blackwell B200 GPU has 208 billion processors. And Nvidia’s GB 200 Grace Blackwell Superchip, combines two of the B200 chips.
What does the future hold?
Well, in a few years time there will be a demand shift in the market for AI semiconductors. Today, Nvidia’s chips are being used for machine learning training. But by, say, 2028, computing speed will become less important. AI will focus on predictions and solving tasks, inference. Less expensive chips will be used for inference. But as costs fall, demand will accelerate. Nvidia will benefit from the so-called Jevons paradox; prices fall but demand increases exponentially to more than offset lower prices.
REDUCING SANCTIONS ON RUSSIA WOULD HURT THE US ENERGY INDUSTRY
But the company also sees strong growth opportunities in robotics, in full self-driving vehicles and its leading role in the development of quantum computing. Top line: Nvidia’s total addressable markets offer revenue opportunities measured in the trillions.
It’s not all rosy. It is always important to remember that export restrictions on chips sold to China are a constant headline risk for Nvidia. Regardless, it is highly probable that Nvidia will be a winner for several more years because demand in the U.S. and across the free world will more than offset lost business in China.
The author owns stock in Nvidia.
James Rogan is a former U.S. foreign service officer who later worked in finance and law for 30 years. He writes a daily note on the markets, politics, and society. He can be reached at [email protected].