Generative artificial intelligence (genAI) like ChatGPT has primarily been housed in the large data centers of service providers and enterprises. Companies typically buy access to an AI platform such as Microsoft 365 Copilot when they want to use genAI services. However, cloud-based systems have their drawbacks, including high energy consumption and network connectivity issues. The industry is now shifting its focus to edge devices, such as PCs, tablets, smartphones, and cars, to offload processing from data centers.
Intel’s vice president, Bill Pearson, highlighted the importance of keeping data on-site in the factory or within the country due to cost and privacy concerns. The edge is becoming the fastest-growing compute segment, with more than 50% of enterprise-managed data expected to be created and processed outside of the data center or cloud by 2025.
Intel, AMD, and Nvidia are now producing more dedicated chiplets and neuro-processing units (NPUs) to assist edge-device CPUs and GPUs in executing genAI tasks. Apple is also rumored to be introducing Apple GPT, a version of ChatGPT that resides on the iPhone, at its Worldwide Developers Conference in June.
Expected soon on…
2024-02-05 09:00:04
Link from www.computerworld.com