Why the AI slowdown doesn’t cripple data‑center real estate
Investors are watching the recent calls for a more measured pace of artificial‑intelligence development with a mixture of anxiety and curiosity. The real question isn’t whether AI will pause, but how the broader wave of digital transformation will continue to shape the physical spaces that power the internet.
The AI‑driven boom that still matters
According to a McKinsey forecast, AI could be responsible for roughly 70% of global data‑center capacity demand by 2030. The same study estimates the total capital needed to meet that demand at close to $7 trillion, with real‑estate components alone potentially absorbing $3 trillion over the next five years, as reported by JLL. Those figures illustrate why data‑center REITs such as Digital Realty and Equinix have become marquee names for anyone tracking tech‑linked infrastructure.
A pause, not a halt
When Anthropic, OpenAI and xAI signaled a slowdown in their model‑training workloads, both companies saw their shares dip. Yet Digital Realty’s chief executive, Andrew Power, cautioned against treating the move as an existential threat. “There's tremendous digital transformation happening that is not connected to AI,” he told Property Play, emphasizing that cloud‑computing growth, enterprise IT upgrades and the migration of legacy workloads to the public cloud remain massive demand drivers.
Balancing cloud versus AI labs
Power noted that hyperscalers now face a strategic choice: allocate new rack space to pure cloud services or dedicate it to AI‑specific clusters. In markets where hyperscalers dominate, this tug‑of‑war can tighten supply and push up lease rates. In other regions, especially those with diversified tenant bases, the impact is muted.
Geography matters
Digital Realty’s footprint includes hubs such as Northern Virginia, Dallas, Chicago, Singapore, Tokyo, Frankfurt and Amsterdam. Power warned that the intensity of competition for space varies dramatically across those locations. For example, Northern Virginia—often dubbed “Data Center Alley”—already experiences high occupancy, so any slowdown in AI demand could translate into a modest softening of rent growth, but not a collapse. Conversely, emerging markets like Singapore or Frankfurt, where the cloud‑first agenda is still expanding, may continue to see robust absorption even if AI projects stall.
What investors should watch
- Capacity utilization trends: Look for quarterly reports that separate AI‑related power usage from standard cloud workloads.
- Tenant diversification: REITs with a broader mix of cloud providers, enterprises and hyperscalers are better insulated from sector‑specific volatility.
- Geographic exposure: Regions with excess supply could feel more pressure, while scarcity‑driven markets may sustain premium pricing.
Looking ahead
While a temporary cooling of AI model training may shave a few percentage points off short‑term demand, the underlying digital transformation engine is unlikely to sputter. Companies that can flexibly re‑allocate rack space between AI and conventional cloud services will emerge stronger, and investors should favor those with a global, diversified tenant roster. In short, the AI slowdown is a blip, not a terminal event, for the data‑center real‑estate sector.
Original reporting via Source.