Although this opportunity is quite expensive, as it requires a monthly payment of $37,000 for a configuration consisting of eight of Nvidia’s most advanced A100 or H100 chips connected together, it could potentially accelerate the AI industry’s growth. This growth has resulted in a staggering 77% increase in the value of its shares this year. As a result, the company has become 5 folds more valuable than its long-time competitor, Intel Corp.
Nvidia-based in Santa Clara, already leads the field for artificial intelligence chips and has assisted companies like Microsoft Corp in building large systems for OpenAI’s ChatGPT services. These services answer questions with text that resemble human-like language and generate images from prompts.
At Nvidia’s yearly software developer conference, CEO Jensen Huang shared that Nvidia is collaborating with partners like Oracle Corp to make their powerful DGX supercomputers accessible to everyone through a simple web browser login. These machines can hold as many as 32,000 of Nvidia’s chips. Huang went on to compare this moment to Apple Inc’s opening up of the smartphone market, dubbing it “The iPhone moment of AI.”
In addition to this partnership, Nvidia is also collaborating with Alphabet Inc and Microsoft to offer its supercomputers as a service for creating new AI products. During the conference, Nvidia introduced new chips and software aimed at making products such as chatbots more affordable to operate daily, even after they have been created using supercomputers.
Semiconductors analyst Hans Mosesmann says these products are “years ahead of the competition.” He praises Nvidia’s leadership in AI software, which holds significant weight and accelerates the industry’s growth.
Nvidia is also teaming up with AT&T Inc to streamline the dispatching of trucks, collaborating with quantum computing researchers to accelerate software development, and partnering with Taiwan Semiconductor Manufacturing Co, an industry giant, to hasten chip development, according to CEO Jensen Huang.
Nvidia’s latest rental service, DGX Cloud, offers developers a chance to access tens of thousands of its chips simultaneously. Biotech firm Amgen Inc and software firm ServiceNow Inc are among the first to take advantage of the service, Nvidia announced.
Moreover, Nvidia has launched an AI Foundations service to help companies train their personalized artificial intelligence models. This service is particularly beneficial for major stock image database owners who can use it to avoid legal disputes concerning the copyright of images used to create AI-generated content.
During the conference, Huang also unveiled new technology aimed at expediting the design and production of semiconductors. This software utilizes Nvidia’s advanced chips to speed up a crucial step in the process, which involves transforming the software-based chip design into physical lithography masks that can be used for printing onto silicon. Typically, this process could take up to two weeks using traditional computing chips. However, with Nvidia’s chips and software, the task can now be completed overnight, while consuming only 5 megawatts of electricity, compared to the previous 35 megawatts.
Nvidia is working with ASML Holding, Synopsys Inc, and TSMC to bring this technology to market, with TSMC starting preparations for production in June.
