Industry Leaders Propose Technical Shifts to Curb AI Energy Waste

Technology firms are proposing structural changes to data center operations to combat the rising energy demands of artificial intelligence. SiTime published a blog post arguing that precision timing can minimize synchronization errors in distributed compute systems, preventing processors from idling or repeating work. The company positions this as a part of a shift toward carbon-aware computing, noting that AI-driven climate solutions should not inadvertently increase the environmental burden of digital infrastructure, especially as some U.S. regulators begin limiting large data center developments.

Complementing this approach, Simon Laskaj of Confluent advocates for a transition from traditional batch processing to real-time data streaming. This "data in motion" strategy aims to reduce energy consumption by optimizing compute cycles and minimizing disk I/O. The push for efficiency arrives amid projections that global power demand from data centers could increase by 165% by 2030, according to Goldman Sachs. In Australia, where the data center sector value is predicted to double to $40 billion by 2028, these architectural shifts are presented as critical for the long-term sustainability of AI infrastructure.