Bain & Co: AI Infrastructure Needs to Face $800bn Shortfall
Bain & Company’s latest Global Technology Report underscores the vast investments required to fulfill AI computing needs by the year 2030.
The consultancy estimates a demand of US$2tn in annual revenue to facilitate the expansion of worldwide data centre capacity, projecting a required capacity of 200GW.
Despite factoring in AI-driven savings, Bain identifies a US$800bn shortfall.
AI demand outpaces infrastructure growth
David Crawford, Chairman of Bain’s Global Technology Practice, outlines the magnitude of this challenge.
“If the current scaling laws hold, AI will increasingly strain supply chains globally,” he says.
“By 2030, technology executives will be faced with the challenge of deploying about US$500bn in capital expenditures and finding about US$2tn in new revenue to profitably meet demand.
“Meanwhile, because AI compute demand is outpacing semiconductor efficiency, the trends call for dramatic increases in power supply on grids that have not added capacity for decades.
“Add the arms race dynamic between nations and leading providers and the potential for overbuild and underbuild has never been more challenging to navigate. Working through the potential for innovation, infrastructure, supply shortages and algorithmic gains is critical to navigate the next few years.”
Bain’s report emphasises that AI compute demand is outstripping Moore’s Law more than twofold.
The USA is anticipated to represent half of the 200GW demand, putting pressure not only on financial resources but also on energy infrastructures.
From experimentation to scaling
Even though some enterprises are already benefiting from EBITDA improvements of 10–25% through AI implementations, most businesses are still in the experimentation phase.
Bain indicates that forerunners are advancing into agentic AI, building platforms that enable autonomous workflows across multiple systems.
These systems necessitate data centres designed for high virtualisation, low-latency connections and seamless real-time data access.
The consultancy outlines four maturity stages of agentic AI, ranging from single-task workflows to complex multi-agent constellations.
The intermediary stages, where capital investment and innovation meet, will pose significant demands on data centre infrastructure.
SaaS and sovereign AI pressures
The report further examines disruptions in the SaaS industry.
AI could broaden markets available to SaaS companies, but it will necessitate strategic transformations in data ownership, monetisation and integration. Data centres will be pivotal to enabling SaaS firms to integrate AI into workflows extensively.
Anne Hoecker, Head of Bain’s Global Technology Practice, stresses the geopolitical pressures impacting digital infrastructure.
“Sovereign AI capabilities are increasingly seen as a strategic advantage on par with economic and military strength,” she says. “While sovereign AI is a global priority, individual countries' goals vary. Therefore, for most countries, achieving full-stack independence is not feasible, at least not today. Considering these differences, global AI standards are unlikely to converge.
“To succeed, multinational firms will need to localise not just compliance, but also their technology architecture. Businesses need to make decisions with optionality, moving boldly where confidence is high and prioritising flexibility where uncertainty rules.”
The fragmentation of semiconductor supply chains adds complexity, with the US and China spearheading separation efforts, while other countries strive to maintain competitiveness alongside sovereignty.
Beyond AI: quantum and robotics
Bain's report also looks at adjacent technologies. Quantum computing holds the potential to generate up to US$250bn in value for industries like pharmaceuticals, logistics and finance, although fully fault-tolerant devices remain a distant future prospect.
The realm of humanoid robotics is also witnessing noteworthy investments, yet Bain warns that most deployments are still in nascent stages and significantly reliant on human oversight.
Implications for investors
Despite challenges, Bain highlights that private equity interest in technology stays robust, though deal activity decelerated in the latter half of 2025.
Investors persist in seeing data centres and AI infrastructure as vital areas for expansion, even though the increasing capital demands of projects will necessitate deliberate funding and strategic partnerships.
The report indicates that meeting AI’s escalating compute needs will require collaborative efforts between tech providers, governments, investors and utilities.
Absent new revenue and capital influx, AI-driven economic ambitions might outstrip the power provision of data centres intended to underpin them by 2030.
