Date: 7 October 2026 Publication Type: Research Project: MaMeeFarm™
Abstract
Digital infrastructure has traditionally treated electricity as an operating input.
Servers consume power.
Data centres pay electricity bills.
Utilities supply the load.
That relationship is changing.
Artificial intelligence, cloud infrastructure, advanced computing, and other large digital loads are expanding rapidly enough that electricity can no longer always be assumed to be available at the required location, volume, reliability, and time.
The constraint is not simply how much electricity a country generates.
Digital capacity also depends on whether power can be delivered through:
- generation;
- transmission;
- substations;
- transformers;
- distribution infrastructure;
- interconnection agreements; and
- reliable operating capacity.
Current evidence illustrates the scale of the emerging challenge. The International Energy Agency reported in 2026 that more than 2,500 GW of generation, storage, and large-load projects were waiting in grid connection queues globally. At the same time, data centres can often be developed much faster than major new grid infrastructure.
Corporate behaviour is also changing.
On 6 October 2026, Google and Constellation Energy announced power agreements totalling approximately 3.59 GW, including long-duration commitments associated with additional nuclear output and power supply in the PJM region. Constellation also announced substantial investment associated with expanding nuclear output.
The significance extends beyond one power agreement.
Large digital companies are increasingly moving beyond simply purchasing electricity toward securing long-term power supply and, in some cases, supporting the investment required to expand future electricity capacity.
This research examines the conditions under which electricity becomes a strategic constraint on digital growth.
Its central finding is that electricity becomes strategically constraining not merely when total energy is scarce, but when reliable power cannot be created, connected, and delivered at the location and speed required by digital investment.
1. Research Question
When does electricity become a strategic constraint on digital growth, and how do generation capacity, grid access, transmission, firm power, interconnection rules, and infrastructure lead times determine where and how quickly digital infrastructure can expand?
2. Electricity Can Be an Input Without Being a Constraint
Every digital system requires electricity.
But electricity consumption alone does not make electricity strategically constraining.
If additional power can be obtained quickly at predictable cost, electricity functions mainly as an operating input.
A company selects a site.
The facility connects to the grid.
Electricity is purchased.
Operations begin.
Under those conditions, electricity affects operating cost but may not determine whether the investment itself can proceed.
The relationship changes when power cannot be assumed to arrive when required.
3. A Strategic Constraint Determines What Can Be Built
Electricity becomes strategically important when its availability begins to determine:
- whether a project can proceed;
- where it can be located;
- how large it can become;
- when it can begin operation;
- how quickly it can expand; or
- whether additional infrastructure must be created before deployment.
At this point, electricity is no longer merely an expense.
It has become a condition of productive capacity.
4. Digital Growth Is Becoming More Power Intensive
Artificial intelligence is one important driver of this transition.
Training and operating advanced AI systems require large concentrations of computing hardware.
Cloud computing, storage, networking, conventional data-centre services, and other digital workloads also contribute.
The category should therefore remain broader than AI alone.
Data-centre demand and AI-specific demand overlap, but they are not identical.
The relevant structural trend is the emergence of increasingly large and concentrated electrical loads.
5. Electricity Demand Is Rising in Major Digital Markets
The U.S. Energy Information Administration projected in October 2026 that U.S. electricity consumption would increase from 4,195 billion kWh in 2025 to approximately 4,288 billion kWh in 2026 and 4,356 billion kWh in 2027.
Growing demand from AI-related and other data-centre activity is one contributor alongside broader electrification.
The important research point is not simply that electricity consumption is increasing.
It is that new demand can arrive in very large increments at specific locations.
6. Concentrated Loads Create Different Infrastructure Problems
A gradual increase distributed across millions of customers is different from a new facility requesting hundreds of megawatts at one location.
Large digital loads can require:
- substation expansion;
- new transformers;
- new transmission;
- new generation;
- network studies; and
- changes in system planning.
A power system can therefore have sufficient electricity in aggregate while lacking sufficient capacity at the requested site.
7. Generation Capacity Is Not the Same as Deliverable Power
A region may possess adequate generation.
That does not establish that a new data centre can receive the required electricity.
Power must move through a physical network.
Several forms of capacity therefore need to remain distinct.
Generation Capacity
Power that can be produced.
Transmission Capacity
Power that can move across the high-voltage network.
Substation Capacity
Power that can be transformed and routed locally.
Interconnection Capacity
Power that a specific project is authorized and physically able to connect.
Deliverable Site Capacity
Power that can actually reach the intended facility under the required operating conditions.
These are related, but they are not interchangeable.
8. Grid Capacity Is Becoming a Bottleneck
The IEA’s Electricity 2026 identifies grids as an emerging constraint on the connection of new generation, storage, and demand.
More than 2,500 GW of projects were reported as waiting in connection queues worldwide.
The IEA also estimated that annual grid investment would need to rise substantially by 2030 from around USD 400 billion to meet expected electricity demand.
Creating more generation alone therefore does not resolve every infrastructure constraint.
The grid must also be capable of connecting and delivering that electricity.
9. Electricity Becomes Strategic When Infrastructure Timelines Diverge
One of the most important constraints is time.
The IEA reports that new data centres can often be developed in roughly one to three years.
Major new transmission infrastructure can require substantially longer periods.
Some grid-enhancing upgrades can be completed more quickly, but major network expansion generally operates on a slower development cycle than many digital investments.
This creates a structural mismatch:
digital demand can arrive faster than electrical infrastructure can be built.
10. Speed-to-Compute Depends on Speed-to-Power
A company can possess:
- capital;
- processors;
- servers;
- land; and
- customer demand.
But those resources do not create operational digital capacity without electricity.
If electrical connection requires several additional years, the investment remains partially unproductive.
This produces an increasingly important relationship:
speed-to-compute increasingly depends on speed-to-power.
11. Interconnection Is Becoming a Strategic Process
Electricity infrastructure is not accessed automatically.
Large loads can require formal connection processes.
Engineering studies may be required.
Network impacts may need to be assessed.
Upgrade costs may need to be allocated.
Connection agreements must be established.
These processes can become major determinants of project timing.
12. Large Loads Are Changing Regulation
The U.S. Federal Energy Regulatory Commission has established a dedicated proceeding addressing the interconnection of large loads to the interstate transmission system.
The proceeding reflects growing regulatory attention to data centres, manufacturing facilities, and other large electricity users.
This demonstrates that large new loads have become significant enough to influence transmission policy and regulatory design.
13. Speed-to-Power Has Become an Explicit Regulatory Objective
In 2026, FERC took additional action concerning the connection of data centres and other large electricity users.
The Commission explicitly connected its action with improving speed-to-power while maintaining reliability and protecting consumers.
The terminology is significant.
Electricity connection speed is becoming part of industrial and digital deployment policy.
14. The Constraint Is Not Only Physical
A data centre can be delayed even when the physical grid could eventually support it.
Institutional rules also matter.
Possible constraints include:
- connection studies;
- tariffs;
- cost-allocation rules;
- permitting;
- queue management;
- contract structures; and
- regulatory uncertainty.
Electricity can therefore become strategically constraining through both physical and institutional infrastructure.
15. Electricity Constraints Have Several Dimensions
Electricity can constrain digital investment in different ways.
Physical
Generation, transformers, substations, cables, and transmission may be insufficient.
Temporal
Required infrastructure may take longer to develop than the digital facility.
Locational
Available power may not exist where investment wants to locate.
Reliability
The facility may require dependable power beyond simple annual energy availability.
Institutional
Connection rules, tariffs, permitting, and cost allocation can affect timing and project structure.
These constraints can overlap.
A project may face several simultaneously.
16. Electricity Volume and Power Capacity Are Different
A facility may consume a certain number of megawatt-hours over a year.
But its infrastructure also needs sufficient instantaneous power capacity.
This distinguishes:
Energy Volume
Total electricity consumed over time.
from
Power Capacity
The rate at which electricity can be supplied when required.
A region can produce sufficient annual energy while still lacking local capacity to support a large new load at peak operation.
17. Reliability Changes the Value of Electricity
Digital infrastructure often requires high availability.
A facility may not be able to treat frequent power interruptions as normal operation.
The relevant question therefore becomes:
Can the required power be delivered reliably when it is needed?
not merely:
Is sufficient electricity produced over the year?
18. Firmness Matters
Electricity resources differ in their availability profiles.
Some generation is dispatchable.
Some varies with weather.
Storage can shift energy through time.
Transmission can move supply between regions.
Demand flexibility can reduce load during constrained periods.
A reliable digital power system can therefore be assembled from different combinations of:
- generation;
- storage;
- grid capacity;
- backup power; and
- flexible demand.
There is no single required technology.
19. Electricity Constraint Does Not Imply One Preferred Energy Source
The growth of digital infrastructure should not automatically be interpreted as evidence that one generation technology is universally required.
Different regions possess different:
- resources;
- grid structures;
- market designs;
- storage;
- transmission; and
- reliability requirements.
The relevant systems question is:
Which combination can provide dependable power at the location and timescale required?
20. Nuclear Power Is Becoming One Part of Digital Power Strategy
Nuclear energy is attracting attention partly because of its ability to provide large volumes of continuous generation.
On 6 October 2026, Google and Constellation Energy announced agreements totalling approximately 3.59 GW.
The arrangements include long-term commitments associated with additional nuclear output and power supply in the PJM market.
The significance is not that nuclear power represents a universal solution.
It is that digital growth is increasingly associated with large and long-duration electricity commitments.
21. Long-Term Contracts Change the Role of the Buyer
A conventional electricity customer purchases supply from an existing system.
A sufficiently large long-term buyer can play a different economic role.
Long-duration commitments can provide revenue certainty that supports investment.
The Google–Constellation arrangements are associated with substantial planned investment and additional output from existing nuclear facilities.
The buyer does not become the utility.
But demand from a sufficiently large buyer can become part of the investment case for expanding electricity capacity.
22. Digital Companies Can Become Anchor Customers for Energy Investment
This changes the economic relationship between digital demand and energy infrastructure.
Under one conventional arrangement, utilities build capacity and customers purchase electricity from the resulting system.
Large digital loads can create another situation in which long-term demand commitments support investment in additional capacity.
The digital company remains an electricity customer.
But its expected demand can become economically important to the development of future supply.
23. New Generation Can Be Associated With Digital Demand
The pattern extends beyond one agreement.
Black Hills announced on 6 October 2026 plans for approximately $1.8 billion of investment from 2027 through 2029 in new generation associated with supplying a proposed Google data centre in Wyoming.
This provides a direct example of anticipated digital demand being linked with future electricity investment.
24. Procurement Is Becoming Infrastructure Planning
When companies must secure power years before expected operation, electricity procurement changes in character.
The question is no longer only:
How much will the electricity cost?
It increasingly includes:
Will sufficient reliable capacity exist when the facility requires it?
Long-term electricity procurement can therefore overlap with infrastructure planning.
25. Location Begins to Follow Power Availability
When electricity cannot be assumed to be available everywhere, site selection changes.
Digital projects may evaluate:
- land;
- fiber;
- tax conditions;
- latency;
- labor; and
- regulation.
But they increasingly also need to consider:
- available power capacity;
- substation capacity;
- interconnection conditions;
- generation proximity;
- future grid upgrades; and
- reliability.
Electricity availability can therefore influence the geography of digital growth.
26. A Cheap Site Can Be Expensive if Power Arrives Late
Land cost alone does not determine infrastructure economics.
A site may offer inexpensive land while requiring years of electrical upgrades.
Another may cost more but provide much faster access to power.
Once deployment delay is included, the relative economic value of the two sites can change.
A year of delayed digital operation can leave substantial capital underutilized.
The strategic value of electricity therefore includes timing.
27. Grid Access Can Become a Scarce Asset
When many projects seek connection to the same network, access itself can become constrained.
The IEA reports large connection queues and describes regulatory and technical approaches intended to allocate or increase usable grid capacity more effectively.
These include changes to queue requirements, flexible connections, and other mechanisms.
Grid access can therefore become part of competition among infrastructure projects.
28. Not Every Project in a Queue Will Become Real Demand
Connection queues should not be interpreted as guaranteed future load.
Projects can:
- change;
- overlap with other requests;
- lose financing;
- move location; or
- be cancelled.
System planners therefore face uncertainty.
They need to prepare for credible future demand without assuming every proposed project will become operational.
29. Planned Demand Is Not Operational Demand
Several states should remain distinct.
Announced Load
A developer states an expected electricity requirement.
Contracted Load
Agreements for some form of supply or connection exist.
Connected Load
Electrical infrastructure has physically connected the project.
Operational Load
The facility is operating and consuming electricity.
These categories should not be combined.
30. Signed Power Does Not Mean Power Exists Today
The same distinction applies to generation agreements.
A company may sign a long-term power contract associated with future investment.
This does not mean all associated capacity is operational immediately.
Equipment may still need to be installed.
Existing facilities may require upgrades.
Grid work may be necessary.
Regulatory approvals may remain outstanding.
Contracted future capacity is therefore different from current operating supply.
31. Grid Expansion Has Its Own Industrial Supply Chain
Electricity infrastructure depends on physical equipment.
This includes:
- transformers;
- conductors;
- switchgear;
- cables;
- circuit breakers;
- towers; and
- substations.
The IEA reports increasing pressure on important grid-equipment supply chains.
Digital growth therefore indirectly increases demand for industries far outside computing.
32. Digital Expansion Can Become an Industrial Capacity Problem
A data-centre project can be financed and technically designed yet still be delayed because electrical equipment or construction capacity is unavailable.
The effective digital infrastructure supply chain therefore extends into:
- electrical manufacturing;
- construction;
- engineering;
- grid operations; and
- generation development.
Digital infrastructure becomes dependent on broader industrial production capacity.
33. Grid Expansion Is Not the Only Response
Electricity constraints do not always require entirely new transmission corridors.
The IEA identifies approaches capable of increasing the usable capacity of existing networks, including:
- dynamic line rating;
- dynamic transformer rating;
- topology optimization;
- advanced power-flow control;
- storage used as a transmission asset;
- reconductoring; and
- voltage uprating.
Some measures can be implemented more quickly than major new high-voltage transmission projects.
Existing infrastructure can therefore sometimes be used more intensively while longer-term expansion proceeds.
34. Flexible Connections Can Also Increase Access
Flexible or non-firm connection arrangements can potentially allow some projects to connect earlier by accepting limitations during specified system conditions.
This creates an important distinction:
full firm connection
versus
conditional grid access.
The usefulness of such arrangements for digital infrastructure depends on how much operational flexibility particular facilities and workloads can tolerate.
35. Digital Loads Could Become More Flexible
Some computing workloads may be movable through time or between locations.
Others require continuous immediate service.
Training workloads may have different flexibility from real-time inference.
Storage, backup generation, and operational scheduling can also affect facility flexibility.
Digital infrastructure may therefore have some potential to participate in grid management rather than functioning entirely as inflexible load.
But flexibility should not be assumed equally across all digital workloads.
36. Reliability and Flexibility Can Conflict
A data centre may prefer guaranteed electricity at all times.
The grid may prefer the customer to reduce demand during stressed periods.
These objectives can conflict.
Future arrangements may therefore need to balance:
- reliability;
- flexibility;
- price; and
- connection speed.
Electricity contracting increasingly interacts with infrastructure design and operating requirements.
37. Cost Allocation Becomes a Strategic Question
When a large new digital load requires network upgrades, the cost must be allocated.
Possible parties can include:
- the data-centre developer;
- the utility;
- regional grid customers; or
- broader electricity consumers.
FERC’s 2026 attention to large-load integration connects faster connection with reliability, consumer protection, and appropriate cost treatment.
Digital expansion therefore creates governance questions as well as engineering questions.
38. Existing Customers and New Loads Share Infrastructure
Grid investments can benefit multiple users.
But they may also be triggered primarily by one large new customer.
Determining how costs should be divided is therefore complex.
Cost-allocation rules can affect:
- project economics;
- electricity rates;
- site selection; and
- investment timing.
Institutional design can consequently influence the geography of digital development.
39. Electricity Markets Must Adapt to Large Loads
Traditional electricity planning often assumed comparatively gradual demand growth.
Large data-centre projects challenge that assumption.
Individual projects can create substantial changes over relatively short periods.
This can affect:
- planning requirements;
- network upgrades;
- system operations;
- connection processes; and
- cost allocation.
Large digital loads therefore represent a new planning challenge for electricity systems.
40. Electricity Becomes Strategic Before It Becomes Scarce
A power system does not need to experience a national electricity shortage for power to become strategically important to digital investment.
The constraint can appear earlier.
A company may discover that:
- connection takes too long;
- the preferred site lacks capacity;
- sufficiently reliable power is unavailable;
- network upgrades are too expensive; or
- future supply cannot be secured with sufficient certainty.
Electricity becomes strategic when these conditions begin changing investment decisions, even while the wider electricity system continues serving existing customers.
41. Strategic Electricity Is Location Specific
Electricity is not perfectly transferable across geography.
Transmission limits matter.
Congestion matters.
Interconnection rules differ.
Generation resources differ.
A megawatt available in one region may have little practical value to a data centre located elsewhere.
Digital growth therefore increasingly depends on where power is available, not simply how much electricity exists nationally.
42. Electricity Constraints Can Redirect Capital
If one region cannot deliver power quickly enough, investment may move elsewhere.
Electricity infrastructure can therefore influence:
- real-estate value;
- industrial clustering;
- data-centre geography;
- generation investment; and
- network expansion.
The power system becomes one factor directing the physical geography of the digital economy.
43. The Electricity Constraint Can Move
Today’s constraint may be generation.
After generation expands, it may become transmission.
After transmission is built, transformers may become restrictive.
Elsewhere the constraint may be interconnection studies, local distribution, cooling, or construction.
Solving one bottleneck can reveal another.
Electricity should therefore not be treated as one undifferentiated constraint.
The limiting condition can change through time and differ by location.
44. Electricity and Digital Infrastructure Are Becoming More Closely Coordinated
Historically, electricity could often appear relatively late in the conceptual description of digital investment:
a facility is planned and then connected to available electricity infrastructure.
Large digital projects increasingly require power considerations much earlier.
Developers may need to evaluate future load, generation availability, grid connection, network upgrades, reliability, and long-term supply while the digital project itself is still being planned.
Electricity planning therefore moves closer to the beginning of the investment decision.
This does not mean every digital project follows one fixed development sequence.
It means that power availability increasingly influences decisions that previously could be made with greater independence from the electricity system.
45. The Boundary Between the Energy and Technology Sectors Is Changing
Digital companies do not necessarily become utilities.
Utilities do not become technology companies.
But their investment cycles increasingly interact.
Large power commitments from technology companies can support generation investment.
Grid availability can determine digital-project viability.
Data-centre demand can influence transmission and generation planning.
Electricity and digital infrastructure are therefore becoming more tightly coupled systems.
46. Electricity Can Play Different Roles in Digital Investment
Electricity does not have the same strategic meaning under every infrastructure condition.
When adequate power is readily available, price may remain the primary electricity concern.
When power cannot be obtained at the required location or time, availability can begin to determine deployment.
When continued growth requires long-term supply commitments, additional generation, network investment, or location decisions based on future electricity availability, power becomes part of strategic capacity planning.
These conditions can overlap.
They should not be interpreted as a mandatory sequence through which every market or project must pass.
Their research value lies in distinguishing situations where electricity is primarily an operating cost from situations where it begins to determine the feasibility and timing of digital investment.
47. Research Considerations for Electricity-Constrained Digital Growth
Several distinctions are important when evaluating the relationship between digital growth and electricity infrastructure.
Planned digital load is different from operational electricity demand.
Announced megawatts do not establish current consumption.
Generation capacity is different from deliverable site capacity.
Electricity may exist within a region while transmission, substations, transformers, distribution, or interconnection prevent it from reaching a particular project.
Annual electricity volume is different from power capacity.
A system may produce sufficient energy over a year while lacking the instantaneous capacity required by a large digital facility.
Contracted power is different from operating supply.
Long-term agreements can support future infrastructure without establishing that all associated capacity exists today.
Reliability requirements affect usable power.
The electricity needed by a digital facility is not defined only by quantity. Availability and operating reliability can also determine whether the supply is suitable.
Infrastructure timelines matter.
Digital facilities and electrical infrastructure can develop at very different speeds.
Connection queues represent potential demand, not guaranteed future load.
Projects in queues have different probabilities of completion.
Grid access can depend on institutional as well as physical conditions.
Tariffs, permitting, queue rules, cost allocation, and regulatory decisions can influence project timing.
Supply chains can constrain grid expansion.
Transformers, switchgear, cables, construction capacity, and other physical inputs affect how quickly electrical infrastructure can expand.
Flexibility varies by workload and facility.
Some digital loads may tolerate conditional access or temporal adjustment better than others.
The limiting condition can move.
Generation, transmission, interconnection, equipment, reliability, cost, or location can become restrictive at different times.
These considerations do not constitute a universal procedure for evaluating digital electricity constraints.
They identify common distinctions required when interpreting whether electricity is genuinely limiting digital deployment.
48. What Current Evidence Can Establish
Current evidence supports several observations.
Electricity demand is rising in major economies, with data-centre growth among the contributors.
Grid connection queues have reached very large levels globally, and grid development has not always kept pace with proposed generation, storage, and new loads.
Data centres can be developed considerably faster than major new transmission infrastructure.
U.S. regulators are actively examining and reforming processes associated with integrating large loads such as data centres into transmission systems.
Major technology companies are entering long-duration power agreements at multi-gigawatt scale, and some agreements are associated with additional electricity investment.
Utilities are also planning generation investments associated with proposed large digital loads.
Together, these observations support the conclusion that electricity availability is becoming increasingly important to digital deployment strategy.
49. What Current Evidence Does Not Establish
The evidence does not establish that:
- all digital growth will be constrained by electricity;
- all data-centre electricity demand is caused by AI;
- all announced data-centre projects will be completed;
- every region requires the same energy technology;
- a signed electricity agreement represents operational capacity today;
- national electricity scarcity is required for a local grid constraint to exist; or
- long-term power contracts guarantee that associated digital projects will proceed.
These claims require additional evidence.
50. Research Implications
Digital growth increasingly needs to be understood through the interaction of two physical and technological systems.
The computing system includes:
- processors;
- servers;
- networks;
- software; and
- applications.
The electricity system includes:
- generation;
- transmission;
- substations;
- interconnections;
- storage; and
- reliability.
A digital investment becomes operational only when the required parts of both systems are available at the same location and time.
This changes how digital capacity should be evaluated.
Hardware procurement alone is insufficient.
Announced data-centre megawatts are insufficient.
Generation statistics alone are insufficient.
A more relevant question is:
How much reliable electricity can actually be delivered to operational digital infrastructure when it is required?
51. Research Limitations
Digital electricity demand remains uncertain.
AI workloads evolve rapidly.
Efficiency can improve.
Some projects announced today will not be built.
Electricity forecasts depend on economic growth, technology adoption, generation development, and regulatory change.
Grid queues contain projects with different probabilities of completion.
Power agreements can include future capacity rather than immediately available supply.
National statistics can conceal severe local constraints.
For these reasons, forecasts and announced megawatts should not be treated as realized infrastructure.
The research value lies in understanding the structural relationship between electricity availability and digital deployment while preserving the distinction between planned, contracted, connected, and operational capacity.
52. Research Conclusion
Digital growth has always required electricity.
What is changing is the role electricity can play in investment.
When power can be obtained readily, electricity primarily functions as an operating input and cost.
When connection delays determine whether a facility can open, electricity becomes a deployment constraint.
When companies must secure supply years in advance, support additional generation, enter long-duration multi-gigawatt agreements, or choose locations according to future grid capacity, electricity becomes part of strategic infrastructure planning.
The important distinction is not simply whether electricity exists.
It is whether electricity can become deliverable power for the intended project.
Generation without adequate transmission may not create usable site capacity.
Transmission without a viable connection does not energize a project.
A long-term contract associated with future investment does not create operating power today.
A completed data centre without sufficiently reliable electricity does not become productive digital capacity.
This leads to the central research principle:
Electricity becomes a strategic constraint on digital growth when reliable power cannot be created, connected, and delivered at the location and speed required by digital investment.
The constraint is therefore broader than energy scarcity.
It can emerge from:
- time;
- location;
- grid congestion;
- equipment;
- reliability requirements;
- regulation; and
- coordination.
Evidence from 2026 illustrates this transition.
Global connection queues contain very large volumes of proposed generation, storage, and large-load capacity, while major electrical infrastructure can require substantially longer development periods than new data centres.
At the same time, large digital companies are securing electricity through increasingly substantial and long-duration arrangements.
Google’s October 2026 agreements with Constellation total approximately 3.59 GW and are associated with long-term electricity supply and additional nuclear output. Black Hills has separately announced substantial generation investment associated with a proposed Google data centre.
These developments do not establish that electricity will become the dominant constraint everywhere.
They demonstrate something narrower and more important.
Electricity can no longer always be treated as an input that automatically becomes available after a digital investment decision has been made.
For some large projects, power availability is moving upstream into the investment decision itself.
Site selection can depend on it.
Project timing can depend on it.
Infrastructure design can depend on it.
The amount of capital that can become productive can depend on it.
The strategic question for digital growth is therefore expanding beyond:
How much compute can be purchased?
It increasingly includes:
How much reliable power can be created, connected, and delivered quickly enough to make that compute operational?
References
International Energy Agency. (2026). Electricity 2026.
International Energy Agency. (2026). Electricity 2026: Grids.
U.S. Federal Energy Regulatory Commission. (2026). Interconnection of Large Loads to the Interstate Transmission System: Docket No. RM26-4-000.
U.S. Federal Energy Regulatory Commission. (2026). FERC Launches Aggressive Targeted Action to Speed Large Load Integration.
U.S. Energy Information Administration. (2026). Short-Term Energy Outlook.
Reuters. (2026, October 6). Google enters massive 3.6-GW power deal with Constellation Energy.
Reuters. (2026, October 6). Black Hills plans $1.8 billion investment to power Google’s data center.
Reuters. (2026, October 6). US power use to beat record highs in 2026 and 2027 as AI use surges, EIA says.