Gartner report mentions FuriosaAI as Sample Vendor, highlights need for power-efficient AI chips
News
Which metric matters most in Gen AI hardware? TOPS? FLOPS? QPS? GB/s?
No. The key thing is ROI, according to a recent AI industry report by Gartner, the famed analysis firm.
The Gartner publication, “Emerging Tech: Adoption Trends for Energy-Efficient GenAI Semiconductors,” singles out operational costs as a “major force” now shaping how businesses are adopting AI.
“Operational effectiveness and cost reduction will force early adoption of energy-efficient generative AI hardware technology,” Gartner’s analysts conclude. In other words, businesses are focusing on return on their AI hardware investments. (The entire report, which is based on 50+ interviews with 31 providers globally along with 90+ use-case studies, is available to Gartner customers here.)
“Ensure the most cost-efficient product solutions are available by evaluating all the GenAI chip vendors’ offerings, selecting those that provide the best mix of performance, power requirements and cost.”
The report includes FuriosaAI as a Sample Vendor, alongside legacy chip giants like NVIDIA, Intel, Qualcomm and AMD, as well as other startups like Cerebras and SambaNova. Furiosa’s second-gen chip, RNGD (“Renegade”), delivers exceptional power efficiency compared to GPUs and directly addresses concerns about the enormous growth in power consumption for all scales of AI deployment. The chip is designed for inference with advanced large language models and multimodal models, and it’s sampling with customers now.
Gartner notes that the GPUs used for advanced Gen AI consume more than 700 watts, with foundries preparing for 2,000-watt GPUs in the not-too-distant future. The cost of powering these chips at scale is now a major consideration for businesses, Gartner finds. And chipmakers must build products that address both capex and opex concerns.
In its list of recommended actions for the next six to 18 months, the first item is: “Ensure the most cost-efficient product solutions are available by evaluating all the GenAI chip vendors’ offerings, selecting those that provide the best mix of performance, power requirements and cost.”
You can learn about RNGD, Furiosa’s power-efficient chip for inference with cutting edge LLMs and multimodal models: https://furiosa.ai/rngd
RNGD is currently sampling with customers who are planning large scale high performance and energy efficient enterprise deployment to reiterate the energy efficiency
“Gartner, Emerging Tech: Adoption Trends for Energy-Efficient GenAI Semiconductors,” was written by Alan Priestley, Nikita Singh, Menglin Cao, Gaurav Gupta, Joseph Unsworth, Akhil Singh.
Gartner is a registered trademark of Gartner, Inc. and/or its affiliates and is used herein with permission. All rights reserved. Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation.
Gartner research publications consist of the opinions of Gartner’s research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.
