Nvidia CEO Jensen Huang presented a plan to lease a cloud service to create powerful and expensive supercomputers used to develop AI (artificial intelligence) technologies such as ChatGPT. for Business.
According to Reuters on March 21, Mr. Jensen Huang said that access to Nvidia ‘s cloud service will not be cheap – around 37,000 USD / month for 8 flagship A100 or H100 chips integrated together; boom But Nvidia’s offering of the service to more corporations has fueled the AI , which has led to Nvidia stock up 77% this year, making the company about five times more valuable than its longtime rival. Intel.
Nvidia already dominates the field of artificial intelligence chips and helps partners like Microsoft build giant systems for OpenAI . At the annual software developer conference, CEO Jensen Huang said that he is actively working with partner Oracle to provide access to any user to the company’s 32,000 chip DGX supercomputers through Web browser.
Nvidia offers the DGX Cloud service, which gives developers access to thousands of Nvidia chips at once.
“The iPhone AI moment has begun,” Huang said in a recent online speech, referring to how Apple has opened up the market to smartphones. He said Nvidia is also working with Microsoft and Alphabet to provide supercomputing solutions to continue creating new AI products.
Nvidia CEO also revealed new chip technology and software designed to help lower the investment cost of chatbots. Hans Mosesmann, semiconductor analyst at Rosenblatt Securities, said Nvidia’s products will be many years ahead of the competition: “Nvidia’s lead in the software side of AI is not only monumental – it’s also spectacular. is accelerating”.
Nvidia is also working with AT&T to make network data transmission more efficient, alongside quantum computing researchers to speed up software development and TSMC to speed things up. chip development. Nvidia’s new rental service DGX Cloud could give multiple developers access to tens of thousands of its chips at once. Biotech company Amgen and software company ServiceNow have already started using the service.
In particular, besides DGX Cloud, Nvidia has also launched a service called AI Foundations to help companies train AI models in their own custom mode. Owners who already have an image database using AI Foundations will avoid legal questions about the copyright of the images used to create AI content.
Nvidia CEO also announced technology to accelerate design and manufacture of semiconductor devices. The technology that uses Nvidia chips and new software will save time in the manufacturing process, from the chip design on software to the lithographic processing of the coating on the silicon wafer.
Normally, the computation process can take two weeks for a traditional computing chip to complete, but Nvidia says its new technology can handle the task overnight and reduce the amount of electricity used in the task. from 35 megawatts to 5 megawatts, saving businesses both investment and environmental protection. With this breakthrough, Nvidia said the company is working with ASML Holding, Synopsys and TSMC to launch new chips in the market. TSMC will be ready to put the technology into production in June 2023.