Share on facebook
Share on twitter
Share on linkedin
Share on pinterest
Share on whatsapp
23 March (Reuters) 22 March (Reuters) Nvidia Corp (NVDA.O) on Tuesday presented a range of new chips and technologies that the company claimed would increase the speed of processing of ever more complex artificial intelligence algorithms, as well as boost competition against chipmakers in the race to win the lucrative business of data centres.
Graphic chips manufactured by Nvidia (GPU), which initially helped in the gaming industry’s expansion and improvement in video quality, are now the primary chips utilized by businesses for running AI apps. The most modern GPU is, known as the H100, can reduce the processing time by a few days to weeks on specific tasks that require training AI models. According to the company.
They were revealed at NVIDIA’s AI developers conference held on the internet.
“Data centres transform into AI factories that process and refine huge amounts of data to enable users to make smart choices,” said Nvidia Chief Executive Officer Jensen Huang in a statement declaring that the H100 chip will be”the “engine” to power AI technology.
Companies are using AI, also known as machine learning, to serve a range of functions, from suggesting what movies to see to aid to discovering drugs, and it is gaining momentum to become an essential tool for business.
The H100 chip will be made using Taiwan Manufacturing Semiconductor’s (2330. TW) advanced four-nanometer process, containing more than 80 billion transistors. The chip will be available before the end of the third quarter of the season, Nvidia stated.
The H100 could also be used to build Nvidia’s new “Eos” supercomputer, which Nvidia claims will be the most powerful in the world’s artificial intelligence system when it goes live later in the year.
Facebook parent company Meta (FB.O) announced in January that it would build the world’s strongest AI supercomputer over the next year. It will be running at close to five exaflops, and Nvidia announced on Tuesday that the supercomputer would be able to run at over 18 exaflops.
Nvidia also revealed a newly designed processor (CPU) known as”the Grace CPU Superchip that comprises Arm technology. This is the first chip manufactured by Nvidia, which utilizes Arm technology. It will be unveiled in the wake of the announcement that the acquisition of Arm Ltd came to an impasse last month due to issues with regulations.
The Grace CPU Superchip, which is expected to be available during the first quarter of this year, to join two processor chips. It will focus on AI and other tasks that require a lot of processing power.
Numerous firms have linked chips using technology that speeds up data transfers between the chips. In the past week, Apple Inc unveiled its M1 Ultra chip, connected with two M1 Max chips. [nL2N2VB1DI]
Nvidia said that the technology is linked to the NVLink C2C technology revealed on Tuesday.
Nvidia is currently working on its self-driving technologies and is expanding its business. It is rolling out its autonomous car computer technology, “Drive Orin”, in March. It also announced that it is anticipated it will be the case that Chinese electronic vehicle maker BYD Co. (002594. SZ) and the luxury electric car maker Lucid Motors (LCID.O) will use Nvidia Drive to power their new fleets.
Danny Shapiro, Nvidia’s vice-head of automotive, informed the company that there would be $11 billion worth of automotive companies within the “pipeline” in the coming six years. This is an increase from the $8 billion forecasts in the previous year. The increase in the revenue forecast resulted from hardware and the increase in regular revenue generated by Nvidia software, as per Shapiro.
Nvidia the shares were essentially unchanged during midday trading.