AI energy efficiency monitoring ranks low among enterprise users, survey by inference CPU specialists finds

LovabledanielsJanuary 19, 202589 Views

Swimlane survey finds many businesses aren’t keeping on top of AI energy needs
Nearly three quarters are aware of the dramatic energy demands needed to train AI models
Just 13% actively monitor AI energy consumption, which may indicate most used off-premise facilities

As the transition from simple algorithms to advanced models significantly increases energy demands, the adoption of agentic AI, known for its advanced decision-making capabilities, is intensifying concerns over energy consumption, new research has claimed.

A survey by SambaNova Systems, sampling over 2000 business leaders from the United States and Europe, found 70% of business leaders are aware of the substantial energy requirements for training models for AI tools, but only 13% monitor the power consumption of their AI systems.

At the same time, 37.2% of enterprises are facing growing stakeholder pressure to improve energy efficiency, and 42% expect these demands to intensify.

Challenges with AI energy demands

Rising energy costs have become a significant challenge, with 20.3% of businesses identifying them as a pressing issue.

Thankfully, 77.4% of businesses are actively exploring ways to reduce power usage by optimizing their models, adopting energy-efficient hardware, and investing in renewable energy solutions.

However, these efforts are not keeping pace with the rapid expansion of AI systems, leaving many enterprises vulnerable to rising costs and sustainability pressures.

“The findings reveal a stark reality: businesses are rushing to adopt AI, but aren’t prepared to manage its energy impact,” said Rodrigo Liang, SambaNova Systems’ CEO.

“Without a proactive approach to more efficient AI hardware and energy consumption, particularly in the face of increasing demand from AI workflows, we risk undermining the very progress AI promises to deliver,” he added.

“By 2027, my expectation is that more than 90% of leaders will be concerned about the power demands of AI. As businesses integrate AI, addressing energy efficiency and infrastructure readiness will be essential for long-term success.”