Business figures The big winner from the AI boom is Nvidia; Blackwell is now in full production

Related Vendor

Wuxi InfiMotion Technology Co., Ltd.

When Nvidia executives presented the figures for the first quarter of fiscal year 2025 on May 22, 2024, some were quite surprised. It was expected that Nvidia would be one of the major beneficiaries of the AI boom. The fact that the company reported a revenue growth of more than 250 percent, however, is indeed impressive.

The electronics and electrical engineering industry is talking about AI, and if there is a company that benefits from the AI boom in the industrial, but especially in the digital environment of the LLMs and AI data centers, then it is Nvidia - despite all regulatory sanctions.

On May 22, 2024, Nvidia's executives presented the figures for the first quarter of fiscal year 2025, which were so impressive that the company's stock spontaneously crossed the psychologically important threshold of 1,000 US dollars.

The annual count of Nvidia briefly explained: Since a fiscal year at Nvidia ends on January 31 of a year, despite the calendar year 2024 we are already in the Nvidia business year 2025. The company's executives briefly reported the following news to their investors:

• Nvidia's revenue in Q1 of fiscal year 2025 amounts to 26 billion USD, an increase of 262 percent compared to the previous year.

• This revenue is greatly supported by the data center business: this business generated 22.6 billion USD in Q1; an increase of 427 percent compared to the previous year.

• Nvidia reports a profit of 15 billion dollars (an increase of 462 percent compared to the previous year).

• The revenue forecast for the second quarter of FY 2025 is 28 billion dollars.

Blackwell as a building block for even more growth

In March 2024, Jensen Huang, founder and CEO of Nvidia, introduced the next generation of the AI platform, which will follow the currently popular and powerful Grace Hopper platform and, according to Jensen's words, should be 30 times better than Hopper at AI training.

The GB200 Grace Blackwell chip connects two B200 Tensor Core GPUs to the Grace CPU via a die-to-die link. Specifically, it is the Ultra-Low-Power NVLink with a data rate of 900 GByte/s. The Blackwell GPU, developed by Nvidia, is supposed to be capable of calculating models with up to 10 trillion parameters. For this purpose, 208 billion transistors are ready on the chip. Not only is Blackwell supposed to work faster, but also more resource-saving, i.e., it requires less energy than Hopper for the same tasks.

Huang on Nvidia's growth outlook: "We are ready for our next wave of growth. The Blackwell platform is in full production and forms the foundation for generative AI on a trillion-parameter scale. Spectrum-X opens us up a brand new market to bring large-scale AI into pure Ethernet data centers. And NVIDIA NIM is our new software offering that delivers enterprise-grade generative AI that can run on CUDA anywhere - from the cloud to on-prem data centers and RTX AI PCs - through our extensive network of ecosystem partners."

"The next industrial revolution has begun - companies and countries are working with NVIDIA to transform the trillions of dollars' worth of traditional data centers to accelerated computing and to build a new kind of data center - AI factories - to produce a new raw material: artificial intelligence," adds Jensen Huang.