Data center energy demand is forcing a redesign of power grids, and emerging markets like Andhra Pradesh are already building the institutions to manage it

Global spending on data centers could reach $7 trillion by 2030, according to McKinsey. That number is not just a headline for infrastructure investors. It is a planning signal for every VP of Operations, CIO, and energy procurement director whose organization buys power, manages facilities, or operates in regions where grid capacity is already stretched.

McKinsey's analysis, published in June 2026 in partnership with the Financial Times Energy Source Forum, frames the surge in AI-driven data center demand not merely as a strain on existing power systems but as a catalyst forcing a fundamental redesign of how grids are built, financed, and governed. That redesign is now visible in capital markets, in utility boardrooms, and increasingly in emerging markets that are racing to build the institutional infrastructure to absorb it.

AI demand has settled the debate, now comes the structural work

A year ago, executive teams in the energy sector were still debating whether AI power demand was durable. That debate, according to McKinsey, is effectively over. Client conversations have shifted from "will this last" to "how do we build for it." The clearest evidence is capital flows: proposed mergers among major US utilities, private equity pouring into dedicated data center platforms, hyperscalers expanding aggressively into Europe, and Chinese and Southeast Asian operators following the same playbook.

McKinsey identifies three distinct investor groups driving this. Private capital is moving into data center assets and the broader supply chain. Traditional energy players, including utilities and manufacturers, are increasing their exposure. And the large cloud service providers, the hyperscalers, continue to anchor demand at scale in the United States while expanding internationally.

The affordability question is real. McKinsey notes that rising energy costs driven by large loads tend to fall disproportionately on residential and smaller commercial ratepayers, making community concerns and public opposition a live boardroom risk. For enterprise operators, this translates into procurement complexity: power purchase agreements in markets under heavy data center load face new pricing dynamics, interconnection queues are lengthening, and site selection for energy-intensive operations requires a more granular grid-capacity assessment than most teams ran even three years ago.

Andhra Pradesh's 237 GW opportunity is met with a formal coordination institution

On the other side of this demand surge sits a set of markets where the renewable energy potential is massive but the institutional infrastructure to deploy it at scale remains underdeveloped. Andhra Pradesh, a southeastern Indian state, is a precise example. Its mixed renewable energy potential stands at 237 GW, according to the state's Integrated Clean Energy Policy as cited by Climate Policy Initiative. Renewables already account for 50.5% of the state's installed capacity, or roughly 15.98 GW as of March 2026, per data from India's Ministry of New and Renewable Energy. Solar makes up 7.5 GW of that, wind 4.4 GW, and hydro plus biomass the remaining 4.07 GW.

The state has set a net-zero target of 2047, three years ahead of India's national 2070 commitment. India's latest Nationally Determined Contribution, as reported by the Press Information Bureau in 2026, targets 60% of cumulative electric power installed capacity from non-fossil-fuel sources by 2035 and a 47% reduction in GDP emissions intensity from 2005 levels. NITI Aayog's 2026 data places Andhra Pradesh as already ahead of the national renewable share average, but reaching the 2047 target requires a step change in coordination, not just capacity.

To bridge that gap, Andhra Pradesh Power Utilities has jointly established a Centre of Excellence for Energy Transition, or CoE-ET. Climate Policy Initiative, which published an analysis of the institution on August 3, 2026, describes it as a strategic mechanism for coordinating planning among state actors and industry stakeholders and for building technical capabilities within utilities. The CoE-ET is designed to function as a shared resource for the broader southern region of India, not just Andhra Pradesh.

What grid redesign looks like in practice for infrastructure operators

These two developments, McKinsey's $7 trillion data center spending projection and Andhra Pradesh's CoE-ET, connect at a practical level for enterprise operators assessing where to site capacity, procure power, or invest in grid-adjacent infrastructure.

McKinsey's analysis specifically frames data center growth as a driver of innovation across generation technology, grid design, and utility business models, not just an incremental demand source. The pressure that AI load creates on aging grid infrastructure is accelerating conversations that had stalled for years: new transmission models, flexible load agreements, and behind-the-meter generation arrangements. Operators building or expanding energy-intensive facilities now face a grid that is being actively renegotiated around them.

In markets like Andhra Pradesh, the institutional work is running alongside the physical buildout. The CoE-ET model, as analyzed by Climate Policy Initiative, addresses a gap that has slowed renewable deployment in many high-potential markets: the absence of a coordinating body with both technical depth and cross-stakeholder authority. Building that body before the capital wave arrives, rather than after, is the structural bet Andhra Pradesh is making.

What this means for your team

• Reassess power procurement assumptions in any market where data center load growth is concentrated. McKinsey's $7 trillion projection through 2030 signals sustained upward pressure on grid costs and interconnection timelines, which will affect pricing and availability for enterprise buyers competing for the same capacity.
• Evaluate grid-readiness as a site selection criterion. In both developed and emerging markets, the gap between installed renewable capacity and the institutional coordination required to deploy it reliably at scale is now a measurable operational risk, not a background assumption.
• Track state and regional energy governance structures in high-potential markets. Andhra Pradesh's CoE-ET model points to a new class of institutional actor that will shape where clean energy infrastructure investment lands in South Asia. Procurement and supply chain teams with footprints in the region should engage with these bodies early.
• Build flexibility into long-term power purchase agreements. As grid architecture is actively redesigned around AI demand, contract structures that assume stable legacy grid configurations may carry more basis risk than they did two years ago.