Hyperscale providers are at the centre of global digital infrastructure. DC Byte’s 2025 Global Data Centre Index [email wall] estimates that public cloud, social media, and AI workloads account for nearly 70% of global data centre demand. Their build and leasing decisions shape where power networks are reinforced or need alteration/expansion, where national, state or international governments adjust policy, and where investment capital flows.
The infrastructure story for enterprises and hyperscalers involves cloud costs, resilience, access to power grids, data sovereignty, environmental concerns, staffing, and the pace at which AI initiatives are predicted to scale.
Hyperscale strategy shapes risk
McKinsey notes that AI workloads are driving a new wave of high-density data centre builds. Goldman Sachs Research forecasts that data centre power demand could rise by about 165% by 2030, with utilisation rates rising over the next few years.
DC Byte’s data shows that hyperscalers are behind the majority of this growth. The concentration of demand has three important implications for enterprises:
- Pricing power: when hyperscalers compete for scarce power and grid connections, costs will affect cloud pricing and reserved capacity terms.
- Availability and latency: if capacity in your preferred metro is fully pre-committed, new users or those expanding provision may be steered to a secondary region, with implications for latency and, in some cases, compliance.
- Strategic dependence: increased numbers of workloads on hyperscale platforms mean exposure to suppliers’ location, energy, and regulatory decisions increases.
Risk is passed to the customer.
Land and latency, power and policy
Earlier waves of cloud expansion optimised for land price, connectivity, and proximity to enterprise demand. Today, the bottleneck is power.
Northern Virginia – home to roughly a third of the world’s data centres – is a case in point. Dominion Energy, the main utility in the region, has warned that connecting very large loads to the grid can take up to seven years, as demand for DCs and other electrification projects outstrip supply.
This is visible in several mature hubs:
- Vacancy rates below 1% in leading metros signal that new capacity is effectively sold out before it’s built.
- Hyperscalers lock in power and land 24-36 months ahead of delivery, turning early grid access into their competitive advantage.
- Entire campuses can be committed to a single tenant, so less space for multi-tenant colocation.
Therefore, capacity planning assumptions that worked five years ago – such as “we can always add another region in this metro later” – may no longer hold. Power and policy are among the first constraints to consider at board level.
Regional shifts
The report describes a decentralisation of hyperscale infrastructure, with capacity is spreading between and inside countries’ borders, as operators search for power and land.
Americas: Two-speed growth
North America remains the most mature hyperscale region, but runs at two speeds:
- Hubs like Northern Virginia still play large in global connectivity, yet face long lead times.
- The US Southeast – states such as Georgia, North Carolina, and Alabama – is emerging as a fast-growth area, offering cheaper land, proactive utilities, and tax incentives.
Investments by Google, Meta, Microsoft, and AWS in these states show how initial builds can establish local ecosystems of subcontractors, grid upgrades, and skilled workers. These attract additional projects.
Companies seeking capacity can expect greater availability in non-traditional metros and a choice of risk profiles: mature hubs predictability but slower timelines; newer hubs offer speed and incentives but less history and, in some cases, community intolerance of new facilities.
Asia-Pacific
The Asia-Pacific (APAC) region is now one of the fastest-growing hyperscale markets. Initial investments are clustered around Singapore, Hong Kong, Tokyo, and Sydney. As land and power constraints tightened, growth has reached Johor, Jakarta, and Bangkok, plus major Indian metros (where digital adoption and government incentives encourage cloud investment).
For enterprises, APAC location choice is as much about regulatory alignment and supply chain resilience as it is about latency.
Europe, Middle East, Africa
In Europe, growth centred historically on FLAP-D: Frankfurt, London, Amsterdam, Paris, and Dublin. As power, land, and planning tightened, attention has moved to Southern and Central Europe (Milan, Spain, Poland), which offer more land and clearer reform programmes, the Nordics for their abundant renewable energy, and the Middle East and Africa, such as Saudi Arabia, the UEA, South Africa, Nigeria, Kenya, Egypt, and Morocco.
One emblematic project is the planned €4 billion hyperscale campus at a former coal plant in Montereau, France, backed by EDF and OpCore. It will reuse existing grid connections and industrial land.
For pan-EMEA workloads, expect a greater choice of regions, but with very different regulatory and sustainability profiles, and a link between region selection and ESG commitments.
Leasing and expectations
Pre-leasing is an increasing popular model for the big hyperscalers. In many mature hubs, projects are sold out before public announcement or ground is broken. In London, for example, hyperscale lease rates have risen by around 30% this year.
For businesses using hyperscalers, lead times will be longer according to how exacting the demand is for the right platform in the right metro. Unit pricing is more exposed to local supply-demand dynamics and power costs, as fluctuations are often passed from hyperscaler to user. Build-to-suit and AI-optimised capacity can be coupled to a single hyperscaler’s architecture, limiting future multi-cloud options for the end-user.
Policy and sustainability
Government policy is now one of the strongest forces shaping where capacity is available and how quickly projects move from design to going live.
In mature markets, the trend is to tighter environmental and energy standards. In Germany, for example, the Energy Efficiency Act (EnEfG) imposes minimum energy-efficiency levels, a commitment to increasing the share of renewable power use over time, and obligations to reuse waste heat where possible.
The UK is prioritising grid access for projects that are ready to build, thanks to long connection queues and speculative reservations, and the Nordics link tax breaks to measures such as heat reuse and renewable integration.
In the US, state-by-state laws create a localised picture. The southeastern states offer tax exemptions, special electricity rates, and sometimes streamlined permit acquisition – measures deliberately designed to attract hyperscalers.
Southern European countries such as Italy and Poland are attempting to position themselves as overflows from FLAP-D, changing zoning rules and restructuring the power grid.
As a rule of thumb, sustainable cloud regions become available with three to seven years wait, and overall, cloud providers’ ESG policies are localised according to extant laws on renewables and ecological measures like re-use of waste heat.
Skills, supply chains, vendor risk
Behind every multi-megawatt campus are construction workers, electrical engineers, and operations teams. Here, the picture is tightening.
The US needs – according to one estimate – around 439,000 extra workers this year in the construction industry to satisfy existing plans for new and expanded DC provision. The UK has announced £600 million of investment to tackle construction skills shortages, and there’s government acknowledgement of persistent vacancies in the sector.
There are also supply chain pressures on the sector. Longer lead times for important equipment such as transformers, switchgear, and cooling systems create delays and bottlenecks. For large-scale projects, these issues and scarcity of labour increases investors’ risk, making finance more expensive, with costs passed onto customers.
Conclusions
Hyperscale build programmes are not a competition as to who can build the biggest campus, but which providers can secure power, align with local policies, and offer predictable capacity at stable prices.
The sum of the existing issues is the time it takes to actually build new capacity. According to CoreWeave’s latest analysts’ call: “While we are experiencing relentless demand for our platform, data centre developers across the industry are also enduring unprecedented pressure across supply chains. In our case, we are affected by temporary delays related to a third-party data-centre developer who is behind schedule.” (CEO Michael Intrator).
Now is an important moment to bring infrastructure realities to discussions in the board room. AI, cloud strategy, and ESG all play into the mix, and understanding of the local and transnational issues is important. The triptych of power, policy, and people shape the cloud roadmap, and hyperscale commitment to massive DC capacity expansion is dominating the market. Whether the debt-funded AI boom continues or collapses will have the biggest impact on enterprise cloud procurement in the next five years.
See also: Video interview: Why Vultr Is Building the Next Hyperscaler: AI, Sovereignty & the Future of Cloud
(Image source: “Lançamento das Obras do Novo Campus Hyperscale” by Governo do Estado de São Paulo is licensed under CC BY 2.0.)
Want to learn more about Cloud Computing from industry leaders? Check out Cyber Security & Cloud Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information.
CloudTech News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.