Check to the king. Nvidia’s undisputed dominance in the production of advanced microprocessors for artificial intelligence is under threat. For the first time. A threat that comes from close by. Just a few kilometers away from the Santa Clara headquarters of the American company. With a somewhat surprising announcement, Advanced Micro Device (AMD) presented its chip to the world a few weeks ago: MI300X. And the technical characteristics already make it the most powerful in the world. So powerful as to allow the Californian company to already win the first contracts with the software giants: Meta, Microsoft and Oracle want those chips. Fundamental pieces to build industries linked to the development of artificial intelligence.
Why microprocessors are so important in the AI industry
The reason is simple. If Artificial Intelligence is the new gold of the technological industry’s desires, microprocessors are the tools that will help us find it. A rather popular metaphor among industry analysts is that they are like pickaxes, shovels and sieves in the hands of new explorers. Without these tools, AI gold is inaccessible. AI is not an industry in itself. But it is technology capable of transforming all possible industries. The gold is there, in its possible developments. And microprocessors are the enabler of this process. During a conference in Santa Clara last month, Lisa Su, CEO of AMD, held an MI300X between her fingers for a few seconds. The AI appeared for a moment in its physical reality: thin, emerald green and silent. A personal success for Lisa Su.
Best of 2023 The Year of AI. And of capitalism by Arcangelo Rociola 11 December 2023
54 years old, Taiwanese, who arrived in the USA at the age of three after spending her first years in Taipei, she became head of AMD in 2014. She said that she took the path to electronic engineering when she understood how to repair cars as a child brother’s remote controls. She studied at Bronx University in New York, then specialized at MIT in Boston. After IBM, Sony and Toshiba, you arrive at AMD in 2011. First you are called to put the accounts in order. The company was considered corrupt. A machine burns money. She cuts back on non-productive sectors and focuses on microprocessors. Three years later she became CEO. She silently puts the pieces together to make AMD ready for the AI challenge.
Nvidia’s advantage. The (successful) run-up of its competitors
Nvidia was one step ahead though. his CEO, Jansen Huang, had understood early on that AI was the future. Over the years he decided to transform the company that produced the best graphics cards in the world (especially for video games) into a company capable of also producing the best microprocessors for AI. The explosion in 2023 of generative AI (ChatGpt and its sisters, so to speak), rewarded Nvidia, bringing it to a value from 300 billion to over one thousand billion dollars. Nvidia’s competitive advantage seemed difficult to chip away at. Not in such a short time at least. But the demand for Ai is much larger than the production capabilities of a single company. There is a thirst for microprocessors. And space for those who can create the best.
Characters Who is Lisa Su, the microprocessor woman who challenges Nvidia by Gabriella Rocco 11 December 2023
Huang, 60, shares Taiwanese origins with Su. But that’s not the only thing that unites Santa Clara and Taipei. In fact, both Nvidia and AMD are companies without factories. They don’t physically produce the chips, but they design them for other companies to produce. Among these, the main one is based in Taiwan. The Taiwan Semiconductor Manufacturing Company (TSMC). Which in its factories is the custodian of industrial skills that make it an absolute leader in the production of ultra-technological chips. A crucial match. Where the economic implications even risk having to give way in importance to the geopolitical ones. Taiwan has long been the target of eastern China’s sights. The TMSC chips are its silicon shield against Beijing’s aggression. A shield that seems destined to become even more solid.
Beijing is not watching: Huawei’s and Moore’s processors
China is not watching the rest of the world. Huawei has put on the market its Mate Pro 60, a smartphone that has a Kirin900 processor developed on a 7 nanometer chip. The news went a little unnoticed but somehow worries the United States because the embargoes wanted by Trump first and by Biden should have made the development of this technology impossible. But something didn’t go as planned for the USA.
Geopolitics Chris Miller: “The chip war is much more than a fight for technological supremacy” by Alessio Jacona 03 September 2023
As Megha Shrivastava pointed out in an article published in The Diplomat, the production of the Kirin9000s “will push the United States to further broaden the scope of its sanctions and export control measures. As a result, U.S. allies in the semiconductor value chain—South Korea, Taiwan, Germany, and the Netherlands, which have borne the brunt of U.S. policies with revenue losses from the Chinese market—will have an increasingly difficult time collaborating with the United States. ”. But again from China on December 19th another news arrived: the Chinese GPU manufacturer Moore Threads announced the creation of MTT S4000, its latest graphics card for artificial intelligence and data center computing workloads. Moore Threads is also collaborating with several other Chinese companies, including Lenovo. All indicators that China is catching up. And quickly.
Artificial intelligence OpenAI desperately needs chips: the one who benefits is Sam Altman by Pier Luigi Pisa 05 December 2023
Interest in AI goes beyond geopolitical significance. One fact explains it well. Lisa Su during the presentation of her microprocessor in November said she expects the market for chips for artificial intelligence alone to reach 400 billion by 2027. Today the entire microprocessor industry, from refrigerators to smartphones, is worth about 600 billion . Su’s estimate goes well beyond the most optimistic forecasts (Gartner for example predicts market growth of around half in the same period). But you have the merit of photographing the situation: the AI race has just begun and US companies already have the keys to a technology destined to change everything. On her he held her between his fingers in the greenish reflections of her. Her eyes seemed to say, “He is here.”