Infosys Flags Power Supply as a Core Constraint on AI Infrastructure Growth

Infosys has identified power supply as a key constraint on the growth of AI infrastructure through 2030, warning that the pace of electricity-system development may not match rapidly rising AI demand.

The company’s assessment is that efficiency improvements and peak-demand management can release capacity in the near term, but do not eliminate the need for new energy infrastructure. Its proposed response combines demand optimisation with investment in transmission and grid infrastructure, low-carbon generation and energy storage.

The analysis differentiates between workload types. Model training can potentially be shifted in time in response to power-system conditions, while real-time applications generally have less operational flexibility. Infosys therefore recommends adapting SLAs, usage policies and tariff structures to the differing shiftability of customer workloads.

For data center development, the company advocates phased, modular construction. Such an approach could allow operators to adjust deployment plans as AI models, chips and cooling technologies evolve, rather than locking projects into a fixed design and demand profile at the outset.

Infosys also argues that energy suppliers and operators should substantiate capacity commitments through historical performance data, real-time monitoring and contractual guarantees. This could strengthen the importance of verified deliverable power—not simply nominal interconnection availability—in data center site assessment and capacity planning.

For the market, the findings suggest that projects combining flexible load management, modular delivery, storage and contractual risk-sharing mechanisms may become more valuable. This remains an analytical inference from Infosys’s assessment rather than evidence of a specific data center project, capacity commitment or infrastructure build-out.