Microsoft reveals it has invested billions in strengthening Azure infrastructure to enable ChatGPT

Microsoft recently revealed how Microsoft spent billions of dollars to enable OpenAI to make ChatGPT a reality.

The cooperation between Microsoft and OpenAI has attracted much attention recently. In January this year, Microsoft announced that it would invest US$10 billion in OpenAI, the AI ​​startup that created ChatGPT. ChatGPT is an AI chatbot that has taken the internet by storm due to its impressive human-like conversational capabilities.

# However, the cooperation between Microsoft and OpenAI actually started a few years ago. According to Bloomberg, Microsoft had spent "hundreds of millions of dollars" before this year on the computing infrastructure needed to develop ChatGPT.

Bloomberg reported that the money was used to build a large supercomputer for training ChatGPT. In two recent blog posts, Microsoft discussed its work in building AI infrastructure and how it plans to make the system more powerful to support more advanced models.

To build the Azure supercomputer that powers the OpenAI project, Microsoft purchased thousands of GPUs made by Nvidia and wired them together. By connecting all these GPUs, Microsoft can provide the massive computing power OpenAI needs to train increasingly powerful AI models.

For a large manufacturer like Microsoft that can invest billions of dollars in promising startups, "hundreds of millions of dollars" of investment may not seem like much, but this move shows that artificial intelligence has become One of Microsoft's top priorities.

This is obvious, because Microsoft recently revealed that Microsoft is working hard to enhance Azure’s AI capabilities and make it more powerful. Matt Vegas, principal product manager for Azure HPC AI, said that Azure recently upgraded new virtual machines that use Nvidia H100 and A100 Tensor Core GPUs, which are connected to Nvidia's Quantum-2 InfiniBand accelerated network architecture launched last year. Microsoft believes that, This will allow OpenAI and other AI companies using Azure infrastructure to train larger and more complex models.

Eric Boyd, corporate vice president of Microsoft Azure AI, said in a statement that Microsoft foresees the market demand for certain special-purpose clusters that are specifically used for large-scale training workloads, and OpenAI was the first to do so. One of the proof points. "We worked closely together to understand what were the key things they were looking for and what were the key things they needed when building a training environment," he said.

The above is the detailed content of Microsoft reveals it has invested billions in strengthening Azure infrastructure to enable ChatGPT. For more information, please follow other related articles on the PHP Chinese website!