How Nvidia became the largest market capitalization in the world

How Nvidia became the largest market capitalization in the world
How Nvidia became the largest market capitalization in the world

Behind this barbaric acronym hides the engine of the new generation of artificial intelligence, in particular so-called generative AI, that is to say capable of creating text, music, images or video without human intervention, in response to a query in everyday language, like the ChatGPT interface.

The graphics processor (GPU – graphics processing unit) or graphics card functions as a much faster and more powerful calculation unit than a classic microprocessor (CPU). It has become essential for managing in real time the requests of generative AI users, who require processing massive quantities of data.

Nvidia’s flagship product, the H100, is by far the most in demand in the industry and is worth tens of thousands of dollars each.

“GPUs are much harder to find than drugs,” Elon Musk quipped last year at an event organized by the Wall Street Journal.

According to research firm TrendForce, ChatGPT requires around 30,000 GPUs to run.

Google, for its remote computing activity (cloud), is also a major consumer of Nvidia processors, as is Amazon for Amazon Web Services (also cloud) or Microsoft.

In mid-March, Nvidia presented the Blackwell, a family of GPUs successor to the H100, which should be marketed by the end of the year. The company billed it as “the most powerful chip in the world.”

A little history…

In 1993, Jen-Hsun Huang, who today calls himself Jensen Huang, founded, with two partners, Nvidia, which became the first company entirely dedicated to graphics processors, hoping to ride the wave of video games, which were beginning to offer more sophisticated animations and 3D.

Nvidia launched the world’s first GPU in 1999 and, with the internet revolution, developed processors adapted to data processing centers or “data centers”, delocalized infrastructures which allow companies and individuals to access computing power up to here reserved for a handful of researchers and large technology companies.

Nvidia’s two main rivals in the graphics card market are two American groups, AMD (Advanced Micro Devices) and Intel, with much older origins. According to Jon Peddie Research, Nvidia accounted for 88% of standalone GPUs shipped worldwide in the first quarter of 2024.

Nvidia designs its processors, but does not have its own production site, like AMD or Apple. If the group is deliberately evasive about its supply chain, according to several specialists and media, most of its products are manufactured by the Taiwan Semiconductor Manufacturing Company, or TSMC, a world benchmark for semiconductors.

Born in Taiwan, Jensen Huang is an atypical boss, with a direct style, a fan of cold humor, and a relaxed demeanor. Like Apple’s Steve Jobs’ turtleneck, the leather jacket he often wears during Nvidia presentations has become his signature clothing.

The man who, at nine years old, was sent to boarding school by his parents in a hamlet in Kentucky, is a supporter of “tolerance for risk-taking” and of the propensity of a company “to be flexible enough to change trajectory quickly”.

Shortly after the birth of Nvidia, the group gave up a contract with the big name in consoles Sega to reorient itself towards a new processor capable of being competitive on the personal computer (PC) market, coming close, in the process, to the filing for bankruptcy.

With AFP



NEXT Gas prices, DPE, savings plan… What’s changing on July 1, 2024