AI INFRASTRUCTURE
AI power shortfall could reshape chip race: Morgan Stanley
Morgan Stanley estimates US data-centre developers could face a 32 GW electricity shortfall through 2028, potentially favouring large, well-financed companies. Some US demand could move to Asia, including India, but the bank does not quantify India's potential share.
BharatVox ·
US data-centre developers could face a 32-gigawatt electricity shortfall through 2028, making access to power and infrastructure a bigger constraint on AI expansion than chip availability, according to Morgan Stanley research reported by Economic Times Markets on October 17.
The bank's report, titled What Does the Power Shortfall Mean for Semis?, estimates the gap at about 34% of projected power requirements, even after allowing for additional power-generation solutions. Its assessment suggests that buying advanced processors will not necessarily translate into usable computing capacity without electricity, suitable land and operational data centres.
BharatVox Graphic
Projected share of global data-centre capacity additions
Projected share attributable to major technology companies and specialist cloud providers.
The same groups are projected to account for nearly two-thirds of additions.
Their projected share remains above half of global additions.
Power access could favour the largest players
Morgan Stanley argues that companies with strong credit backing, predictable revenues and large electricity requirements are securing power more quickly. That could put smaller developers at a disadvantage as investment in AI infrastructure runs ahead of electricity availability.
The bank estimates that major technology companies and specialist cloud providers could account for about 94% of global data-centre capacity additions in 2026. It projects their share at 64% in 2027 and 57% in 2028, and warns that the concentration of access to power could put other participants under pressure.
Chip orders face a deployment bottleneck
The risk extends to semiconductor suppliers. Morgan Stanley warns that customers could postpone chip deliveries or cancel orders if the data centres intended to house them cannot begin operating.
The bank nevertheless expects Nvidia and Broadcom to meet their 2027 forecasts, citing their visibility into customer demand and infrastructure availability. It sees greater exposure for suppliers of memory chips, optical equipment and other supporting components if deployment slows.
The constraints may last beyond the immediate construction cycle. Morgan Stanley projects that the cumulative US data-centre power deficit could widen to 68 GW by 2029.
Asia could gain demand; India's share remains unspecified
Power constraints could encourage US technology companies to locate more computing infrastructure abroad, including in Asian markets such as India. Morgan Stanley's Asia data-centre team estimates that 14 GW of demand from US hyperscalers could shift to Asia through 2030.
That is a regional projection, not an estimate of capacity coming to India: the report does not specify an Indian share. The bank also cautions that continued international expansion would not resolve the underlying US power shortage.
Fact-check this story · Source: Economic Times Markets: AI race may depend more on access to power, infrastructure than chip supply: Morgan Stanley. This AI-assisted account is based on the linked source, not independent on-the-ground reporting by BharatVox. Any opinion is separately labelled.
Join the conversation (0)
Sign in to share your view on this story.