Major Three AI chip 1: Nvidia, AMD ,Intel: new generation of chips shipped into the Taiwan market were also on the rise in TW, Nov 26 At the consumer electronics event over the weekend, Nvidia (NASDAQ:NVDA) CEO, Jensen Huang said Rubin will be the chip-maker’s next-generation AI chip platform, It is expected to be released in 2026. Rubin: This came after Blackwell in March which has been shipping generic datacenter chips for some time now. The competition that was offered by it was styled as the ultimate – the world’s most powerful chip at the time by Nvidia. Recently, in his speech at the seminars of the National Taipei University, Taipei Huang pointed out that Rubin would also be equipped with new Graphics Processing Units (GPUs) for the same.
Reuters And Advanced Networking Chips Would Be “Perfect” For Launch Plans, Huang Said To A Lecture At National Taiwan University In Taipei.”pch_club : ‘Computing platform in 3 years time, more of it will be done in ARM than in x86 architecture that will be a revolution,’ he said before Computex, an annual show held in Taipei, begins. During this year’s GTC, the statement made is that ” It will take accelerate computing and artificial intelligence to define the future. He commenced running a new cycle for the semis on the “one year beat”.Increased interest rate in firms using generative AI has boosted the share prices of chip stocks.
By 2020, the market leader Nvidia quadrupled its share price. Richard Windsor, founder of Radio Free Mobile, an independent research company focusing on the digital and mobile ecosystem, said: Moore of the same financial firm has put together this chart which sums up Nvidia’s control of the AI semiconductors with the firm claiming a 70% market share in the current generation. Though, there is rivalry as they are also interested in outselling Nvidia with AMD (AMD) and Intel (INTC) soon introducing new products. Speaking at a Taipei event on Monday, AMD chief executive Lisa Su then presented the firm’s new AI chips, and in the next two years, lots of new products are expected to surface. The next generation of its MI325X accelerator will be rolled out in Q4, she added.
A day later, Intel’s CEO Patrick Gelsinger unveiled the sixth generation of its Xeon chips to be used in data centres and its Gaudi 3, AI accelerator chips. Of the two, the latter that goes head to head with Nvidia’s H100 is a third cheaper according to Zhang.
“Top Priority: Leading the Charge in Innovation”
Since the USA and China are now vying to one-up the other as to whose nation has the capacity to bring to life the most applications of generative AI, an influx of new orders for the data center chips on which these programs run is increasing the demand for the chips. Nvidia and AMD – both headed by Taiwan-born American CEOs who relocated to America as kids and hence were part of the same generation of the same family – were previously familiar to gamers mainly as the sellers of the GPUs through which every game came to life with virtually real graphics.
Although the two companies do continue to compete, to some degree, in this area, they also compete in the generative AI space to drive new, more advanced systems like ChatGPT.AI remains the number one priority for us and the industry is ready to move through a period that will be extremely interesting, Su added. We debuted last year at the top if the inference performance charts and for memory size and compute with the MI300X, he added that availability is now on an annual basis, a new family of products every year, she said.
This new chip will be the next generation of the MI300 that will feature more memory along with higher memory bandwidth and overall compute capability, Su mentioned. They expected in 2025 for MI350 and then MI400 the following year and each year it will repeat for every new family.