The end of data centres? Apple and Meta’s bet against the cloud | AI Decoded

Data centers are increasingly becoming political flashpoints due to their immense energy, water, and land requirements, alongside concerns about their role as strategic military targets. The emergence of open-weight AI models, such as the recent collaboration between Meta and Apple, offers a potential alternative by enabling AI to run locally on consumer hardware rather than relying solely on cloud-based infrastructure. While this shift promises greater privacy and reduced reliance on centralized servers, high hardware costs currently limit these models to enterprise and developer use. Experts argue that while local models may address some logistical and privacy concerns, they do not eliminate the existential risks posed by frontier AI. Public opposition to data center expansion reflects a broader societal anxiety regarding the rapid, often uncontrolled, trajectory of AI development and the lack of transparency in its long-term impact on the global economy.