During APRICOT 2025, Billy Lee Kok Chi, the chairman of CSF Advisers, highlighted the significant challenges associated with repurposing various types of real estate into AI datacenters. Lee examined the feasibility of different facilities, providing statistics on compliance likelihood. He warned that multi-story commercial buildings are almost entirely incompatible, while urban lands present high costs. However, large industrial factories and expansive rural plots show more promise under certain conditions, with existing datacenters being the best candidates for conversion.
Lee noted that multi-story commercial buildings have a 99 percent chance of noncompliance when repurposing them for AI datacenters because structural limitations hinder necessary features.
He emphasized the high land and construction costs associated with repurposing empty land in urban centers, resulting in an 80 percent likelihood of infeasibility for successful AI datacenter development.
Collection
[
|
...
]