A stark gap in readiness is emerging as enterprises race to integrate artificial intelligence (AI) into their operations, with new research highlighting widespread concern about the energy demands of AI systems alongside a significant lack of action to address the issue. According to a survey commissioned by SambaNova Systems, nearly half of business leaders, 49.8 per cent, are concerned about the growing power requirements of AI, yet only 13 per cent actively monitor the energy consumption of their systems.
This disconnect comes at a critical juncture. The rise of advanced AI workflows, particularly Agentic AI, is expected to drive power consumption to unprecedented levels. While 72.4 per cent of respondents acknowledge the significant energy required to train large AI models, fewer than 60 per cent are aware of the energy-intensive nature of AI inference, the process of running trained models to generate outputs. As inference workloads increasingly dominate AI usage, this lack of awareness poses a significant challenge for operational scalability.
Rodrigo Liang, CEO of SambaNova Systems, underscored the urgency of the issue: “Businesses are rushing to adopt AI but aren’t prepared to manage its energy impact,” he said. “Without proactive strategies for efficiency, the rising power demands of AI risk undermining its potential benefits. By 2027, I expect energy consumption to become a key performance indicator tracked by corporate boards.”
The research also reveals a growing recognition of energy efficiency as a strategic priority. Over 56 per cent of respondents believe energy efficiency will play a crucial role in future planning, driven by cost management and sustainability pressures. However, the current rate of action falls short. Among organisations that have widely deployed AI, only 77.4 per cent are taking steps to reduce power usage, with strategies including hardware and software optimisation (40.4 per cent), adopting energy-efficient processors (39.3 per cent), and investing in renewable energy (34.9 per cent). These measures, while promising, remain insufficient given the rapid pace of AI adoption.
The introduction of Agentic AI, which enables systems to make autonomous decisions and adapt to real-world environments, is amplifying these challenges. As businesses scale AI deployment, they are facing rising power costs, with 20.3 per cent of surveyed companies identifying these costs as a pressing issue. Additionally, 37.2 per cent report increasing stakeholder pressure to improve AI energy efficiency, with a further 42 per cent anticipating similar demands in the near future.
The shift toward more sustainable AI systems is likely to reshape the landscape of AI hardware. Current GPU-based solutions, while powerful, are criticized for their excessive energy consumption and associated costs. As power demands grow, enterprises are expected to turn to more efficient alternatives that can deliver high performance without unsustainable energy requirements.
The broader implications of these findings point to the need for education and strategic planning to bridge the awareness gap. For many organisations, addressing the energy impact of AI is not just an operational necessity but also a commercial imperative. Rising energy costs and the scalability challenges of power-intensive AI workflows could hinder long-term adoption if left unchecked.
Liang emphasised the critical role of energy efficiency in AI’s future: “The rapid pace of AI adoption underscores a critical need for enterprises to align their strategies with the power requirements of AI deployment,” he added. “Addressing energy efficiency and infrastructure readiness is essential for long-term success.”
The research points to a pivotal moment for enterprises embracing AI. Those that prioritize energy-efficient technologies and align their infrastructure strategies with the demands of advanced AI workflows will be better positioned to navigate the challenges ahead. As the industry scales, the need for sustainable, cost-effective solutions will only grow, placing energy efficiency at the heart of AI’s evolution.
