Nvidia launches AI supercomputer, entering the data center playing field

In addition, the company also revealed new processors and technologies that it believes will enhance the computing power of increasingly complex artificial intelligence algorithms. Specifically, the new graphics chips (GPUs), the “heart” of the AI ​​infrastructure: H100 and Grace CPU Superchip based on ARM technology. This is the first ARM-based processor since Nvidia’s failed acquisition of this chip company.

“Data centers are becoming AI factories — processing and refining mountains of data to extract important information,” said Jensen Huang, Nvidia CEO, during a webinar for developers. AI developer, while calling H100 the “engine” of the AI ​​platform

Nvidia says the new technologies will work together to reduce compute time from weeks to days for some jobs related to training AI models.

Companies are using AI and machine-learning in many areas, from recommending the next video on TV and phones to researching new pharmaceuticals.

“It’s clear that Nvidia’s latest statements show that it is becoming a threat to Intel and AMD in the cloud and data center markets,” said Bob O’Donnell, chief analyst at Microsoft. TECHnalysis Research, commented.

Intel is the largest chip maker for data centers, but this fertile field is showing increasing competition.

Vlad Galabov, Head of Data and Cloud Applications Research at research firm Omdia, said he is concerned about the power consumption of the H100 and that it may reduce the appeal of this chip line on the market. market.

Colette Kress, Nvidia’s CFO, says that the new processors help drive AI, and the company’s market opportunity is estimated at $1 trillion, from gaming consoles to enterprise systems.

One of the driving forces behind companies using Nvidia’s processors is the company’s use of open source software. However, Nvidia said the company will aim to commercialize the software business more in the future.

“Our software business has reached several hundred million dollars and we believe growth opportunities are still there,” Kress said, adding that the software business will help Nvidia make up for the increased costs from the shortfall. chip components supply.

The Nvidia executive also said that software for the automotive market will also be a key driver in the future: “This area is becoming our next billion-dollar business.”

This month, the company is expected to deliver shipments of “Drive Orin”, a chip for autonomous vehicles, to customer BYD, the Chinese electric vehicle maker. Lucid Motors, the maker of luxury electric vehicles, is also expected to use Nvidia Drive in its next-generation products.

Vinh Ngo(According to Reuters)