Global Hyperscale Data Center Capacity
Global Hyperscale Data Center Capacity
Experts expect the world's hyperscale data centres to need 92 GW by 2030 — two and a half times their 2024 capacity.
Hover to highlight a group, click to show or hide itTap a group to show or hide it
Our Analysis
How do we interpret these results
A typical nuclear power plant in the U.S. produces one gigawatt (GW) of power, and the IEA estimates that the total installed capacity of hyperscale data centers globally in 2024 reached 36 GW. Experts, the public, and superforecasters expect the total installed capacity of hyperscale data centers to be in line with IEA projections of 62-108 GW by 2030: The median expert’s central forecast is 92 GW of installed capacity by the end of 2030, a 2.6x increase from 2024. Other groups give slightly lower forecasts: the public predicts 83 GW and superforecasters 86 GW. Experts’ and superforecasters’ central forecasts for 2030, however, are statistically indistinguishable.
Forecasts by group
Median forecasts from each group
| group | 2026 | 2027 | 2030 |
|---|---|---|---|
| Experts | 50 GW | 62 GW | 92 GW |
| Superforecasters | 48 GW | 60 GW | 86 GW |
| Public | 46.2 GW | 57 GW | 83 GW |
Click any cell to compareTap any cell to compare
* Denotes non-overlapping bootstrapped confidence intervals between group aggregates.
Rationale analysis
The main lines of reasoning driving the spread in estimates
Physical constraints
The most significant consideration involves the extent to which physical limitations—power, land and water availability, wait times to connect to the grid, supply chain bottlenecks, etc.—will cap growth. Low-forecast respondents tend to expect physical constraints in the data center build-out, despite the level of capital investment and demand for compute. Of the U.S., one writes, “Converting capital to operational capacity faces significant friction: interconnection queues stretch 5+ years, PJM’s [largest regional transmission organization in the US] grid becomes capacity-constrained by summer 2026, and regional operators are delaying transmission upgrades.” High-forecast respondents typically acknowledge these and other constraints, but believe that unprecedented capital allocation will overcome them enough to allow for rapid growth, particularly if we see “strategic sovereign investment in faster-building jurisdictions (e.g., UAE, China).”
Interpretation of IEA projections
The IEA report linked to this question contains four different hyperscale growth scenarios. Low-forecast respondents typically judge the report’s Headwinds case, which envisions slower-than-anticipated AI adoption and bottleneck issues, to be more credible whereas high-forecast respondents tend to be more confident “that AI will be economically worth this investment" and that therefore the Base Case (which is anchored by robust-growth industry projections) or the Lift Off case (which envisions more-rapid-than-anticipated AI adoption and flexibility in data-center siting) is more likely to materialize.
Durability of demand
Many high-forecast respondents argue that short-term growth is effectively guaranteed by committed capital: “Major growth is baked in now until 2027…” By 2030, one predicts “AI will be demonstrating clear value for the economy and national security” and another that as “AI moves from chatbots to compute-intensive agents that run continuous loops,” demand for inference will surge. Low-forecast respondents are more skeptical. In the near-term several highlight that “the IEA notes that roughly 20% of planned data-center additions in the Base Case are at risk of delay due to grid bottlenecks and siting constraints” and that in the long-term, there is the risk of a market correction or simply diminishing returns. One draws a parallel to the dotcom era, predicting “infrastructure will be overdeveloped and we will have a high data-center capacity growth rate during the boom years…but face slowing development once the enthusiasm for burning cash has vanished.”
Impact of efficiency gains
Low-forecast respondents frequently highlight the potential for efficiency gains to limit data center growth. (This potential is also captured in the IEA’s High Efficiency scenario.) “Over time AI energy need will saturate because of more efficient hardware,” wrote one, and another that “AI's energy consumption could decrease dramatically through algorithmic improvements.” Still another noted that “people like Ilya [Sutskever] believe the pre-training era is ending and we are entering a research driven phase again, with room for growth in RL based fine tuning that does not necessarily require exponential compute.” Most high-forecast respondents don’t dismiss these arguments, but instead tend to believe that “efficiency gains partially offset the need for infinite capacity expansion but do not fully solve the energy burden.”
Public backlash
Several low-forecast respondents emphasized the potential for a backlash to data center growth, particularly in Western democracies, to stymie growth more than it already does. “With news of diesel powered data centers right next to towns and residential energy bills set to jump as a result of these energy hungry data centers…something will have to give,” wrote one, and another that “environmental regulations and other onerous construction obstacles" in the West will hamstring development. Many high-forecast respondents, however, point to deregulatory trends: “[The] current US administration is likely to minimize regulations on the AI industry and advocate for resources to maintain competitive levels with China.”
Rationale examples
High-quality rationales from forecasters across the range
High-forecast respondents
I anchor on the IEA’s estimate that hyperscale installed capacity was ~36 GW in 2024 (up from ~20 GW in 2020) and their scenario projections that put 2030 hyperscale around ~62-108 GW depending on conditions.That implies roughly ~10-20%/yr growth through 2030, so extrapolating gives around ~50 GW by end-2026 and high-50s by end-2027 as a central path, with uncertainty around whether the buildout is front-loaded or delayed by interconnection/transformers/permitting. I keep a fat upside tail because the capex and construction pipeline is enormous (e.g., multi-GW of builds starting/completing annually in mainstream forecasts), and AI demand remains a strong driver.
All major analyses (IEA, Goldman Sachs, McKinsey) seem to converge on a scenario where global hyperscale data center installed capacity nearly triples by 2030, reaching between 120 and 130 GW.
I assumed growth would stay steep through the late 2020s because hyperscalers are racing to expand AI training and inference capacity, with massive capital budgets already locked in through 2026–27. The jump toward 50–60 GW around 2026–27 reflects committed buildouts from Amazon, Google, and Microsoft, as well as new entrants like Oracle and regional AI campuses in the Middle East and Asia. By 2030, I expect demand for power-hungry model training and specialized chip clusters to keep capacity climbing toward roughly 70–80 GW, but not hitting the IEA’s most aggressive projections because of physical bottlenecks, grid strain, transformer shortages, and permitting slowdowns.
I’ve softened my stance on a ‘bubble’ since Wave 3; I now view short-term volatility less as a systemic crash and more as a rotation where hype in generative text is offset by durable growth in ‘physical’ AI (drones, AVs) and specialized verticals like drug discovery. Consequently, I expect infrastructure demand to remain robust, buffered by strategic sovereign investment in faster-building jurisdictions (e.g., UAE, China) that decouples buildout from Western stock tickers. Ultimately, my bullish 2030 median is predicated on an ‘Agentic Shift’: as AI moves from chatbots to compute-intensive agents that run continuous loops, the demand for inference will explode, requiring a massive expansion of installed capacity.
Low-forecast respondents
There are many conflicting forces at play, and many questions that [will] determine the outcome: will the investment climate shift (i.e. “the bubble burst”)? If so, when?; will there be innovations that lead to a shift from compute-intense LLMs?; how will innovations affect AI efficiency in the future?; how rapidly will AI be adopted, i.e. the demand for hyperscalers grow?; how strong will the opposition to hyperscalers, e.g. by local communities, be and how will politicians and powerful tech companies react?; how quickly can hyperscalers be built and be ready to operate?; can the production of powerful AI chips keep up? What if China invades Taiwan?... To me, the “downside” pressures seem more dominant, despite the current volume of investments in compute.
I am basing my answer on conversations I have had with Emirati strategic planners as they go all-in on AI. Basically, while AI is the rage globally, constructing the generation capability to support is easier said than done. Europeans and Americans talk about it but will be hamstrung with environmental regulations and other onerous construction obstacles; states like the UAE don’t have this. The basic point is only a fraction of the countries that talk about this can actually get it done, which is why my numbers remain on the low side.
Grid interconnection queues, permitting, and long lead times for transformers, switchgear, and other electrical gear can delay projects, and the IEA explicitly flags that a meaningful share of additions could be at risk of delay if bottlenecks are not addressed.
Data Centers require capital, land, permission, connectivity, power, water, and hardware. The choices for land are constrained by the availability of permission, connectivity, power and water. Land furthermore comes in chunks that must be acquired in sufficient quantity before work can proceed on the other required resources. Capacity therefore comes in chunks gated by the acquisition of all the parts. The Headwinds case appears to be the most likely case.