NVIDIA and Microsoft Have Announced to Create The World's Most Powerful AI Cloud-based Supercomputer
NVIDIA and Microsoft have announced a multi-year collaboration that will create the world's most powerful AI supercomputer. The cloud-based supercomputer will operate on GPUs and will incorporate a full stack of AI software. It will have the capability to process large amounts of data at high speed.
The Massive Cloud AI computer will use GPUs to train the massive AI model. The company has recently closed a $65 million funding round for the project. The two companies are utilizing NVIDIA's latest A100 GPUs to train the giant AI model. The GPUs will be paired with the best-of-breed RoCE network powered by NVIDIA.
The massive cloud AI computer will use thousands of GPUs and will run on Microsoft's Azure cloud. The system will cost around $12,000 and will be powered by tens of thousands of Nvidia's A100 and H100 chips. The A100 chip can cost upwards of $12,000, and the H100 chip can cost more than $120,000.
NVIDIA is also building a virtual reality platform, dubbed the Omniverse. This platform will run digital twins that are large and complex. It will also run software that enables designers to collaborate with 3D models. The system will also have storage capacity for data.
As the AI industry grows and matures, its economics will make it more affordable for developers to deploy AI applications in the cloud. With these capabilities, developers can build high-performance AI applications using GPUs. These applications are available across multiple platforms, and can be deployed on various NVIDIA GPU platforms.
Adept, a machine learning research lab, is using NVIDIA and Oracle Cloud Infrastructure to create a massive cloud AI computer. The project will run thousands of NVIDIA GPUs on clusters of OCI bare-metal compute instances. The technology will allow the team to train massive AI models faster. The goal is to make it possible for knowledge workers to collaborate more effectively with the assistance of artificial intelligence.
The technology required to create massive AI systems is massive and requires vast amounts of data. It also requires talented engineers and massive computing power. These are all components Nvidia is building today. Microsoft is also partnering with Nvidia to create this massive cloud AI computer.
The MT-NLG model developed by Microsoft and Nvidia has unmatched accuracy in natural language tasks, including text prediction, reading comprehension, and common sense reasoning. However, the ongoing problem of bias in AI continues to plague the field. One of Microsoft's chatbots turned misogynistic within a day. The model picks up stereotypes from the data it consumes.
The company's growth has been remarkable, with more than 70 percent share in the GPU sector. With the expansion of the company into new markets, Nvidia's market cap is $50 billion, with a trailing earnings multiple of 40 times. The company's success has helped its executives amass a fortune of $2.4 billion.