Enterprises Urged to Make Smarter Choices to Cut GenAI Emissions

Generative AI (GenAI) carbon emissions has captured the imagination of enterprises worldwide, offering transformative capabilities across industries. However, alongside its promise, concerns are mounting over the environmental toll of its development and deployment, particularly the carbon emissions associated with training and running large language models (LLMs).

Experts argue that businesses need to make better, more informed choices to reduce the environmental impact of GenAI. A recent study has shed light on the staggering carbon emissions tied to the training and inference of these models, urging enterprises to take actionable steps to curb their environmental footprint.

The study, titled Toward Sustainable GenAI using Generation Directives for Carbon-Friendly Large Language Model Inference, highlights the potential for cutting emissions by over 40% through strategic choices in model inference. The report emphasizes that the industry’s current trajectory may lead to unsustainable energy use unless proactive measures are adopted. Researchers advocate for implementing generation directives—methods to guide how LLMs process and output information—to lower the carbon intensity of operations without sacrificing functionality.

The urgency to address this issue is compounded by the rapid adoption of GenAI carbon emissions across businesses. According to research by Enterprise Strategy Group (ESG), 42% of organizations have already implemented GenAI technologies within their operations, with an additional 43% actively exploring or planning deployment. While the technology promises improved efficiency and innovation, its energy demands remain a pressing concern.

Industry analysts warn that the environmental cost of GenAI could undermine its long-term benefits unless businesses adopt strategies to reduce emissions. Recommendations include optimizing model efficiency, using renewable energy for data centers, and implementing carbon-aware computing to schedule operations during times of low carbon intensity on the energy grid. Some enterprises have already started taking steps to offset their carbon emissions, aligning with broader sustainability goals.

Despite these challenges, GenAI’s integration into the business world continues to grow. From enhancing customer service through AI-driven chatbots to optimizing supply chain operations, companies are embracing the technology as a competitive advantage. However, without a deliberate effort to address the associated energy use, businesses risk exacerbating their carbon footprints at a time when sustainability is a global priority.

To address these challenges, stakeholders across the technology sector are calling for greater collaboration between enterprises, researchers, and policymakers. By pooling resources and expertise, the industry could establish standardized practices for sustainable AI development and deployment. In addition to that, the adoption of energy-efficient hardware and software solutions could further reduce the environmental impact of GenAI.

Some tech giants have already set an example by committing to carbon-neutral or carbon-negative goals in their operations. Microsoft, for instance, has pledged to be carbon-negative by 2030, with a focus on powering its data centers—key enablers of AI systems—with renewable energy sources. Similarly, Google continues to invest heavily in clean energy initiatives to support its AI infrastructure.

The conversation around GenAI emissions extends beyond technological solutions. It also involves rethinking how organizations prioritize sustainability within their broader corporate strategies. For many companies, the challenge lies in balancing the demand for cutting-edge AI capabilities with the imperative to reduce carbon emissions. As organizations increasingly integrate AI into their workflows, addressing these competing demands will require a shift in mindset and operational practices.

The findings of the recent study and the growing awareness of GenAI’s environmental impact underscore the need for immediate action. While GenAI offers unprecedented opportunities for innovation and efficiency, its promise must not come at the expense of sustainability. As enterprises continue to explore the possibilities of AI, they must also take responsibility for mitigating its environmental impact, ensuring that the technology’s future aligns with a sustainable vision for the planet.