At the SC24 supercomputing conference, Intel revealed "Jaguar Shores," its next-generation AI inference chips. These chips are expected to succeed "Falcon Shores," which are slated for a 2025 release. Both Jaguar Shores and Falcon Shores are likely to be GPUs designed for AI inference, potentially incorporating a hybrid GPU-ASIC design.
Intel's "Jaguar Shores" will likely leverage the company's in-house 18A manufacturing node, taking advantage of technologies like RibbonFET and backside power delivery for enhanced efficiency. This comes as Intel continues to invest in its AI portfolio, complementing its existing x86 offerings with enterprise-focused, cost-efficient inference solutions.
"Our AI investments will complement and leverage our x86 franchise – with a focus on enterprise, cost-efficient inferencing. Our roadmap for Falcon Shores remains," - Intel Spokesperson
With "Jaguar Shores," Intel aims to compete against NVIDIA and AMD, who currently dominate the AI training and inference markets. Intel is primarily targeting the inference market with these upcoming chips.
"We continuously evaluate our roadmap...we are committed to providing superior enterprise AI solutions across our CPU and accelerator/GPU portfolio," - Intel Spokesperson.
This announcement comes amidst leadership changes and roadmap adjustments at Intel, including the cancellation of the Rialto Bridge GPU and a shift in the Falcon Shores design from a hybrid CPU+GPU to a dedicated GPU. The company is facing challenges in the client CPU market and is striving to maintain competitiveness in the rapidly evolving AI landscape.
Source: HPCWire
