Intel's AI chip plans seem to have included a new product, as a new report claims that a GPU with low-power requirements is in development.
Intel's 'Low-Power' AI Chip Could Target Inference Workloads & An Approach Similar To Qualcomm's Inference Cards
Well, Intel hasn't been in the best of positions in the AI industry, given that the company has a limited compute portfolio compared to some of its competitors. Intel has relied on its Gaudi lineup for several years now, but based on what we have seen until now, little progress has been made in market adoption. A report by CRN revealed that Intel plans to release an AI chip with low power consumption alongside Jaguar Shores, which will likely target the inference markets.
Outside of Jaguar Shores, the company has an additional, unannounced GPU design with a lower power requirement for servers on its road map that could arrive next year at some point, according to a source familiar with the company’s plans.
The report hasn't disclosed the specifics of Intel's new product, apart from the fact that it is an unannounced GPU design that will debut alongside JGS. Of course, such a chip would likely be targeted towards inferencing workloads, since Jaguar Shores will pan out to be a high-end release intended for training by the firm, which will also go up to rack-scale solutions. Now, we have some good guesses about what this GPU design could be, and here's how it could roll out.
Our best guess for this 'unannounced' design would be a Battlemage-based SKU intended for Edge AI workloads, and Intel has already disclosed intentions to release such a model in the past. Team Blue is known to be working on the 'BMG-G31' silicon and intends to integrate up to 24 GB VRAM in one of the models; hence, this could be one of the options. However, since the chip is expected to debut by next year, Battlemage could become obsolete by then; hence, a solution might come under another lineup, but right now, this is all speculation.
It seems like 'low-power' AI-focused devices might be a segment Intel could target next year, and they could play a role similar to how Qualcomm's Cloud AI 100 inference cards work out, which could be easily deployed on consumer systems.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
