AI BriefingQ&A: ASUS Reveals Partnership with FuriosaAI
Key point
ASUS and FuriosaAI are strengthening collaboration on next-generation chip and server design to address the AI inference market.
Details
FuriosaAI and ASUS unveiled their strategic partnership at MWC Barcelona 2024, showcasing the Gen 1 Vision NPU for computer vision inference. The two companies are working closely together on the design and manufacturing of a second-generation chip for LLM (Large Language Model) inference, slated for release this year.
The data center industry is currently undergoing a structural shift from AI training-centric workloads to AI inference-centric ones. With the spread of generative AI, there is a surging need for specialized server designs tailored to workload characteristics—compute-bound, I/O-bound, memory-bound—as well as heterogeneous computing.
Key challenges in the AI server market include:
- Thermal management: the rise of liquid cooling technology and greater efficiency in air-cooled systems
- Scalability and cost: addressed through innovative data center designs and ecosystem partnerships
- Performance optimization: meeting requirements around latency, power consumption, reliability, and software
ASUS has been collaborating with FuriosaAI by exchanging design feedback from the early stages of product development. For example, to improve the thermal performance of the Gen 1 NPU, the two companies worked closely together, with ASUS proposing modifications to the heat sink, enhancing overall product completeness through this tight co-development process.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.