The continuous incorporation of AI into corporate strategies has spurred advancements in computing power, resulting in new technological developments in digital infrastructures.

Large Language Models (LLMs), such as those developed by OpenAI and Google, have attracted considerable attention for their impressive capabilities in natural language understanding and generation, powering a range of applications from business chatbots to complex data analysis tools.

However, LLMs are known for being resource-intensive, consuming significant amounts of energy and water.

For instance, a query to ChatGPT can consume up to 10 times the electricity of a standard Google search.

Training these models also necessitates data centers that require vast quantities of water for cooling purposes.

The training of GPT-3, for example, was estimated to consume 1,287 MWh of electricity, equivalent to the annual energy usage of 120 American homes.

Microsoft’s operations illustrate this trend, showing a 34% increase in water consumption during 2022, primarily due to AI activities.

In response to these resource-heavy models, Small Language Models (SLMs) have emerged as a practical alternative, providing similar functionality with a lower environmental impact.

Read the complete story in the August 2025 edition of Sustainability Magazine.

These models, with parameters ranging from millions to 10 billion, demonstrate significant efficiency in energy, memory, and storage usage.

In contrast to massive LLMs, SLMs utilize transformer architectures optimized through methods like knowledge distillation and quantization, allowing them to deliver task-specific performance with minimal resources, making them suitable for specific applications such as custom email summarization and customer service solutions.

What are SLMs?

SLMs, unlike their larger counterparts, provide notable advantages in sustainability efforts.