From Ideation to Implementation: How Microsoft Achieved its ChatGPT Goals

Microsoft Corp MSFT pledged to build a massive, cutting-edge supercomputer for OpenAI in 2019 when it invested $1 billion.
Microsoft had to find ways to string together tens of thousands of Nvidia Corp's NVDA A100 graphics chips, the workhorse for training AI models, and change how it positions servers on racks, Bloomberg reports.
OpenAI needed access to robust cloud computing services for long periods as it tried to train an increasingly large set of AI programs called models.
Microsoft executive vice president Scott Guthrie said it cost Microsoft over several hundred million dollars.
"We built a system architecture that could operate and be reliable at a very large scale. That's what resulted in ChatGPT being possible," said Nidhi Chappell, Microsoft general manager.
"That's one model that came out of it. There's going to be many, many others."
Microsoft uses the same resources it built for OpenAI to train and run its own large AI models, including the new Bing search bot.
Microsoft is already at work on the next generation of the AI supercomputer, part of an expanded deal with OpenAI in which Microsoft added $10 billion.
Training a massive AI model requires a large pool of connected graphics processing units in one place, like the AI supercomputer Microsoft assembled.
For the AI supercomputer, Microsoft had to develop software to get the best results from the GPUs and the networking equipment.
Microsoft also deploys graphics chips for inference which are geographically dispersed throughout the company's more than 60 regions of data centers.
The company is adding the latest Nvidia graphics chip for AI workloads, the H100, and the newest version of Nvidia's Infiniband networking technology to share data even faster.
Price Action: MSFT shares traded higher by 2.61% at $255.09 on the last check Monday.

Market News and Data brought to you by Benzinga APIs

Comments

Popular Channels

Tools & Features

Partners & Contributors

About Benzinga