95% of AI Projects Fail. The Data Centre Boom Is Stalling. One Problem Is Behind Both.

David Gal from Samsara highlights that most of the construction industry has not yet begun the foundational step that precedes artificial intelligence adoption. According to Gal, three quarters of firms remain at stage zero of a four-stage digital journey—still relying on manual methods like clipboards and phone calls to track equipment. This lack of digitisation has real costs; one firm lost twenty million dollars worth of equipment in a single year without awareness. The financial impact of not digitising is significant but often invisible.

A KPMG survey of over two thousand senior executives across twenty countries revealed nearly a third are unsure about the sources of their increasing AI expenses. Complementing this, research from MIT found that 95% of corporate AI projects fail—not due to technological issues but because of inadequate organisational groundwork. While the AI technology itself performs, the surrounding structures and processes often do not support successful implementation.

The ongoing boom in data centre construction, which has helped maintain strong order books despite a flat housing market, is beginning to slow. This is not due to funding shortages—major technology companies are investing over six hundred and fifty billion dollars this year—but because of supply chain delays. Key components like transformers and switchgear now have lead times extended from two years to five, creating bottlenecks. Even with land, capital, and chips secured, physical construction is being held back.

The obstacle behind these three challenges—construction’s digital readiness, AI project failures, and data centre supply chain delays—is the same. The construction industry, often labelled backward, is uniquely positioned to understand and address this shared problem. Listeners are encouraged to join the conversation by sharing their thoughts on this issue in the comments of this week’s LinkedIn post.