According to CNBC reports, Microsoft in the recently released earnings report, for the first time, will be related to the GPU added to the risk factors of data center services appear disruptions.
▲ image source Pexels
In the earnings report, Microsoft said, “The growth of our data center business is dependent on buildable land permits and predictable energy, network and server supply, including the availability of graphics processing units (GPUs) and other components.”
That’s one of three paragraphs in the earnings report that mentions GPUs, but Microsoft didn’t mention them once in last year’s report. similar statements haven’t appeared in recent earnings reports from other major tech companies, either.
As part of a complex partnership, OpenAI relies on Microsoft’s Azure cloud service to run ChatGPT and various AI model calculations. Microsoft has also begun using OpenAI’s models to augment existing products such as Outlook, Word, and Bing with generative AI.
These efforts and interest in ChatGPT have led Microsoft to seek out more GPUs than it had anticipated, and it has begun to seek out more GPU support outside of its own data centers after signing a deal with NVIDIA-powered CoreWeave, which rents out GPUs as a cloud service to third-party developers.
Meanwhile, Microsoft has spent years building its own custom AI chips, and as it was previously reported, the focus on ChatGPT has prompted Microsoft to accelerate the deployment of its chip business, with Google, Amazon, and Meta all releasing their own AI chips.
Microsoft CFO Amy Hood previously told analysts on a conference call that Microsoft expects to continue to increase capital expenditures this quarter to pay for data centers, networking hardware, CPUs and GPUs.