According to CNN, the increasing demand for artificial intelligence (AI) development inadvertently exposes the limits of the global chip supply chain.
Microsoft’s annual report highlights a persistent chip shortage as a risk factor for investors. services The company said it will continue to identify and evaluate opportunities to expand data center locations and increase server capacity to meet growing customer needs, particularly for AI. Microsoft data centers depend on licensed land, power, network supplies, and servers, including GPUs and other components.
According to industry analysts, the chip crisis has hit businesses large and small, leaving the AI industry unlikely to make big strides for at least a year or more. The shortfall affects companies building AI products and tools and indirectly affects businesses and end users who hope to adopt the technology for their own purposes
At a hearing before the US Senate Judiciary Committee on May 16, OpenAI CEO Sam Altman said that as few people as possible use ChatGPT because chatbots are having a hard time responding to requests made by users due to a lack of GPU. However, a representative of OpenAI committed that the company still ensures enough storage for users.
This problem has many people thinking of shortages during the pandemic that has caused gaming enthusiasts to pay high prices for computer graphics cards and equipment. At the time, production delays, labor shortages, global shipping disruptions, and constant competitive demand from cryptocurrency miners made GPU supply scarce.
Moody’s senior vice president of services, Raj Joshi, said that no one can model how quickly and to what extent demand for chips will grow. Meanwhile, AMD CEO Lisa Su said there are many customers interested in AI solutions, but the company has a lot of work to do. AMD is expected to launch AI chips later this year to compete with Nvidia.
Last year, the US passed the Chips and Science Act to fund billions of dollars in the domestic chip industry and chip research and development, but those investments are not specifically targeted at boosting GPU production.
According to experts, the chip shortage is expected to ease as more manufacturers and Nvidia’s competitors expand their offerings, however, this could take up to 2-3 years.
The CEO of AI startup d-Matrix Sid Sheth said companies can use AI models that are small, easy to train, and less computationally expensive, or develop new ways of computing to be independent. too much on traditional CPUs and GPUs.