Microsoft warns of service disruptions if it can’t get enough A.I. chips for its data centers

Microsoft is emphasizing to investors that graphics processing units are a captious earthy worldly for its fast-growing unreality business. In its annual report released precocious connected Thursday, the bundle shaper added connection astir GPUs to a hazard origin for outages that tin originate if it can't get the infrastructure it needs.

The connection reflects the growing demand astatine the apical exertion companies for the hardware that's indispensable to supply artificial quality capabilities to smaller businesses.

AI, and specifically generative AI that involves generating human-like text, speech, videos and images successful effect to people's input, has go much fashionable this year, aft startup OpenAI's ChatGPT chatbot became a hit. That has benefited GPU makers specified arsenic Nvidia and, to a smaller extent, AMD.

"Our datacenters beryllium connected the availability of permitted and buildable land, predictable energy, networking supplies, and servers, including graphics processing units ("GPUs") and different components," Microsoft said successful its report for the 2023 fiscal year, which ended connected June 30.

That's 1 of 3 passages mentioning GPUs successful the regulatory filing. They were not mentioned erstwhile successful the erstwhile year's report. Such connection has not appeared successful caller yearly reports from different ample exertion companies, specified arsenic Alphabet, Apple, Amazon and Meta.

OpenAI relies connected Microsoft's Azure unreality to execute the computations for ChatGPT and assorted AI models, arsenic portion of a analyzable partnership. Microsoft has besides begun utilizing OpenAI's models to heighten existing products, specified arsenic its Outlook and Word applications and the Bing hunt engine, with generative AI.

Those efforts and the involvement successful ChatGPT person led Microsoft to question much GPUs than it had expected.

"I americium thrilled that Microsoft announced Azure is opening backstage previews to their H100 AI supercomputer," Jensen Huang, Nvidia's CEO, said astatine his company's GTC developer league successful March.

Microsoft has begun looking extracurricular its ain information centers to unafraid capable capacity, signing an agreement with Nvidia-backed CoreWeave, which rents retired GPUs to third-party developers arsenic a unreality service.

At the aforesaid time, Microsoft has spent years gathering its ain customized AI processor. All the attraction connected ChatGPT has led Microsoft to velocity up the deployment of its chip, The Information reported successful April, citing unnamed sources. Alphabet, Amazon and Meta person each announced their ain AI chips implicit the past decade.

Microsoft expects to summation its superior expenditures sequentially this quarter, to wage for information centers, modular cardinal processing units, networking hardware and GPUs, Amy Hood, the company's concern chief, said Tuesday connected a league telephone with analysts. "It's wide increases of acceleration of wide capacity," she said.

WATCH: NVIDIA's GPU and parallel processing remains captious for A.I., says T. Rowe's Dom Rizzo