HomeArtificial IntelligenceArtificial Intelligence NewsMicrosoft and OpenAI are Planning a $100 billion Supercomputer

Microsoft and OpenAI are Planning a $100 billion Supercomputer

According to a report, Microsoft and OpenAI are collaborating to construct an artificial intelligence supercomputer known as “Stargate,” which might cost up to $100 billion.

Citing a person who discussed the matter with Sam Altman and another who has seen preliminary cost estimates, the publication announced the estimated $100 billion cost. Additionally, it appears likely that Microsoft would fund the project, which might only be completed by 2028 and cost up to 100 times as much as some of the largest data sets.

According to the story, Stargate will be the largest of several supercomputers that Microsoft and OpenAI intend to construct over the course of the following six years. Stargate is part of the fifth and final phase of the supercomputer plan, which is now in the middle of the third phase. According to the analysis, obtaining AI chips would account for a considerable portion of the costs for the next phases.

The lack of graphics processing units, or GPUs, which are needed to crunch data for AI models, is still a bit of a bottleneck in the AI development process. The surge in generative AI led to an explosive increase in the demand for these processors. Furthermore, due to the production of some of the most advanced AI-capable chips, one manufacturer, Nvidia, virtually controls the industry. Additionally, it is finding it difficult to fulfill orders despite rising demand. Furthermore, the cost of its chips is only rising.

Nvidia’s CEO Jensen Huang revealed that the Blackwell B200, the company’s newest AI processor, will retail for between $30,000 and $40,000 per unit shortly after the company’s GTC 2024 event. The chip beats all others on the market by a significant margin, and the business has invested over $10 billion in its research and development.

However, a number of tech firms, including Google Cloud, Samsung, Arm, Intel, and Qualcomm, have united to form the “United Acceleration Foundation,” an open standard accelerator programming paradigm that seeks to undermine Nvidia’s dominance in AI hardware and software.

Source link

Most Popular