Data‑Center Shortfalls Threaten Price Stability
Alicia Garcia Herrero, Natixis’ chief APAC economist, warned on Bloomberg Television that the surge in artificial‑intelligence deployments could lift prices if the construction of data‑center capacity cannot keep pace with the exploding demand for compute power. The comments were made during a live interview on October 26, 2026, as companies across Asia accelerate AI‑driven projects.
Herrero explained that AI models now require orders of magnitude more processing than traditional software, pushing firms to rent or build massive server farms. If supply lags, the cost of electricity, hardware and cloud services will rise, feeding through supply chains and potentially stoking broader inflation. She noted that while AI promises efficiency gains, the short‑run pressure on energy grids and semiconductor inventories could outweigh those benefits.
The economist highlighted that global data‑center capacity is expanding at roughly 10 percent annually, yet AI workloads are growing at double‑digit rates each quarter. „We are seeing a mismatch,” Herrero said. „If the gap widens, firms will bid up the price of compute, and that cost will be passed on to consumers.” She cited recent spikes in power tariffs in Singapore and South Korea, where utilities have raised rates to fund new high‑density cooling systems.
Could AI Adoption Accelerate Inflation Faster Than Expected?
In addition, the supply chain for advanced chips remains tight. The ongoing semiconductor shortage, exacerbated by geopolitical tensions, limits the ability of manufacturers to meet the surge in GPU and ASIC demand. Herrero warned that without coordinated investment from governments and the private sector, the AI boom could become a new source of cost‑push inflation, especially in economies already grappling with tight labor markets.
Analysts have long debated whether AI would be deflationary by automating tasks and lowering production costs. Herrero’s view flips that narrative, arguing that the transitional phase may be inflationary. „The paradox is that technology meant to reduce costs can, in the interim, raise them,” she noted.
She pointed to early adopters in the financial services sector, where AI‑driven risk models demand real‑time data processing. The resulting need for on‑premise compute clusters has driven up leasing rates for rack space by up to 30 percent in the past six months. Similar trends appear in manufacturing, where AI‑controlled robotics require low‑latency connections, prompting firms to invest in edge‑computing facilities.
Outlook and Policy Implications
If the supply‑side constraints persist, central banks may face renewed inflationary pressures, complicating monetary policy at a time when many are already tightening rates. Herrero urged policymakers to consider incentives for renewable‑energy‑powered data centres and to streamline permitting processes for new facilities. She also suggested that industry consortia could share excess compute capacity, smoothing demand peaks.
In the longer term, once the infrastructure catches up, AI is expected to deliver the promised productivity gains. However, the next 12 to 18 months could see price volatility in sectors heavily reliant on cloud services, from streaming media to e‑commerce. Stakeholders are advised to monitor data‑center capacity metrics and energy pricing trends closely.
Frequently Asked Questions
Why would AI increase inflation now? AI workloads demand far more compute power than traditional applications, straining data‑center capacity, raising electricity and hardware costs, which businesses then pass on to consumers.
What can governments do to ease the pressure? They can offer tax breaks for renewable‑energy data centres, fast‑track construction permits, and support research into more energy‑efficient chips.
Will the inflationary effect be temporary? Most experts expect the surge to subside once new data‑center capacity is built and supply chains normalize, allowing AI’s efficiency benefits to dominate.