A significant influx of AI funding has primarily benefitted Nvidia, making it the most valuable chipmaker in the world.
According to Nvidia Corp. CEO Jensen Huang, the company intends to improve its AI accelerators year. In 2025, a Blackwell Ultra chip will be released, and in 2026, a next-generation platform known as Rubin will be developed.
On the eve of the Computex trade show in Taiwan, the company – which is now best known for its artificial intelligence data center systems – also unveiled new tools and software models. During a keynote speech at National Taiwan University, Nvidia’s CEO stated that the company views the emergence of generative AI as a new industrial revolution and anticipates being heavily involved when the technology moves to personal computers.
A significant influx of AI funding has primarily benefitted Nvidia, making it the most valuable chipmaker in the world. But now that a large portion of its revenue comes from a small number of massive cloud computing companies, it wants to grow its clientele. Huang anticipates that as part of the expansion, a wider range of businesses and governmental organizations – from pharmaceutical companies to shipbuilders – will adopt AI. He revisited ideas he introduced in the same location a year prior, such as the notion that people devoid of AI capabilities will be left behind.
According to Huang, “We are seeing computation inflation”. Huang stated that traditional computing methods are unable to keep up with the exponential growth in data that has to be processed. The only way that costs can be reduced is by utilizing Nvidia’s accelerated computing methodology. He claimed that Nvidia’s technology resulted in 98% cost savings and 97% less energy consumption, calling it “CEO math, which is not accurate, but it is correct.”
Huang stated that HBM4, the next generation of high-bandwidth memory that is becoming a manufacturing bottleneck for AI accelerators, with leader SK Hynix Inc., mostly sold out until 2025, will be used in the next Rubin AI platform. Other than that, he did not provide particular details about the next goods, which will come after Blackwell.
Nvidia’s history of selling gaming cards for desktop PCs comes into play as computer manufacturers strive to include more artificial intelligence (AI) features into their devices.
Using Computex, Microsoft Corp. and its hardware partners are showcasing new laptops with AI enhancements under the Copilot+ brand. Most of those gadgets that will hit the market are built on a new processor type from rival to Nvidia, Qualcomm Inc., which will allow them to run longer between battery charges.
Even while the devices work well for basic AI tasks, Nvidia claims that adding a graphics card will significantly boost their performance and enable new capabilities for well-known applications like gaming. According to the firm, these machines are available from PC manufacturers including Asustek Computer Inc.
Pretrained AI models and technologies are being made available by Nvidia to assist software developers in adding more new features to the PC. They will take care of complicated jobs like choosing whether to process data locally on the system or transfer it via the internet to a data center.
Nvidia is also introducing a new architecture for server PCs that use its technology. Businesses like Dell Technologies Inc. and Hewlett Packard Enterprise Co. use the MGX program to expedite the launch of their products for usage by businesses and government organizations. Rivals AMD and Intel Corp. are also utilizing the design to their advantage, putting their CPUs next to Nvidia chips in servers.
As per the business, previously disclosed solutions like Nvidia Inference Microservices (NIM), also dubbed “AI in a box” by Huang, and Spectrum X for networking are now widely used and publicly available. Additionally, free access to the NIM products will be provided. Without having to worry about the underlying technology, businesses may launch AI services more quickly with the help of a set of intermediary software and models called microservices. Businesses that use them are then required to pay a use fee to Nvidia.
Additionally, Huang advocated for the deployment of digital twins in a simulation known as the Ominverse by Nvidia. He demonstrated the scale achievable by demonstrating Earth 2, a digital twin of the planet Earth, and how it can be used to carry out more intricate jobs like modelling complex weather patterns. He pointed out that the instruments are being used by Taiwan-based contract manufacturers like Hon Hai Precision Industry Co., popularly known as Foxconn, to plan and run their plants more effectively.