For an AI data center, access to electricity can determine whether a site works, how quickly it can open, and how far it can expand.
A utility connection is not simply a line item in the development process.
It may require:
Solar Harmonics helps California data center developers evaluate the grid environment surrounding a project and identify where utility infrastructure, solar, battery storage,e and other energy resources may fit into a broader path to power.
The objective is not to work around the utility.
It is to understand the grid requirements early enough to make better development decisions.
Those questions should be addressed before grid access becomes the constraint that controls the project.
Data center site selection is often discussed in terms of land, fiber, zoning, water, taxes, and proximity to major markets.
For AI infrastructure, one additional factor can outweigh nearly all of them:
Can the local electricity system support the load?
A data center may require tens or hundreds of megawatts of new capacity.
That demand can be far greater than the electrical load originally planned for the surrounding area.
The result is that a technically attractive site may still face substantial power constraints.
The first question is not how much electricity exists somewhere on the grid.
It is how much usable capacity can be delivered to the specific site.
That can depend on:
Capacity that appears available at a regional level may not necessarily be available at the point where the data center needs it.
A site near major transmission infrastructure may appear well positioned, but proximity alone does not guarantee usable capacity.
The project may still require:
Each requirement can affect cost and schedule.
The grid strategy should also consider what happens after the first phase is energized.
A campus may begin with one building and eventually grow into several.
If the local grid can support the initial load but not the full build-out, the development team may encounter the same constraint again during expansion.
That is why Solar Harmonics looks beyond the first connection.
The real question is:
Is there a credible power path from initial energization to the project’s ultimate load?
Interconnection is often discussed as if it were a single approval.
In reality, connecting a large AI data center to the electric system can involve a sequence of technical, commercial, and construction decisions.
The process varies by utility and project, but may include:
Early conversations with the serving utility or municipal power provider can help establish whether the project is within the range of what the local system may be able to support.
At this stage, the development team may begin to understand:
These early discussions can be critical during site selection.
A utility may need to study how the new data center load affects the existing electric system.
That analysis may consider:
The findings can change the economics and schedule of the project.
The interconnection process may identify infrastructure that must be added or expanded before the data center can receive its full load.
Potential requirements can include:
The larger the required load, the more significant these requirements may become.
One of the most important questions is who is responsible for the infrastructure required to serve the project.
Depending on the utility, tariff, and project structure, costs may be allocated in different ways.
Developers need to understand:
These issues can materially affect site economics.
Approval does not mean electricity is immediately available.
After studies and agreements are complete, the required infrastructure still has to be designed, permitted, procured and constructed.
That is where long equipment lead times and utility construction schedules can become critical.
The interconnection path affects:
It should therefore be evaluated alongside the broader development strategy, not treated as a utility task that happens in the background.
The earlier the grid path is understood, the more options the project team has to respond to it.
Solar Harmonics begins by looking at the grid environment surrounding the project.
The objective is to identify the questions, constraints,s and energy opportunities that should be understood before major development decisions are finalized.
The process begins with the data center itself.
That may include:
- Initial electrical demand
- Future electrical demand
- Development phases
- Desired energization dates
- Reliability requirements
- Redundancy requirements
- Expansion plans
This establishes what the grid ultimately needs to support.
Next, we look at the local power environment.
That may include:
- Investor-owned utility
- Municipal power provider
- Community Choice Aggregator
- Transmission system
- Distribution system
- Nearby substations
- Existing infrastructure
The purpose is to understand who controls each part of the power path.
Potential questions may include:
- How much capacity appears to be available?
- Which substations could serve the project?
- Are there known transmission constraints?
- Is new distribution infrastructure required?
- What utility upgrades may be necessary?
- Are major loads already committed nearby?
-
This helps identify where the largest uncertainties may exist.
The next step is to understand what the serving power provider may require.
That can include:
- Applications
- Studies
- Deposits
- Engineering
- Service agreements
- Infrastructure design
- Construction
Testing
- Energization
The specific process will vary by utility and project.
The project may have one or several limiting factors.
Potential constraints can include:
-
Substation capacity
- Transmission capacity
- Distribution capacity
- Equipment lead times
- Utility construction
- Land requirements
- Interconnection timing
- Cost responsibility
The goal is to identify which factors are most likely to control the project schedule.
If the grid path is constrained, Solar Harmonics can also evaluate whether other resources may deserve consideration as part of the broader energy strategy.
These may include:
- Solar
- Battery storage
- Distributed energy
- Offsite renewable generation
- Alternative procurement
- Phased load growth
These resources are not assumed to eliminate utility requirements.
They are evaluated for the flexibility they may add.
The Solar Harmonics approach also looks beyond the data center property.
Potential local energy opportunities may exist across:
- Warehouses
- Industrial properties
- Commercial buildings
- Municipal facilities
- Parking areas
- Suitable open land
Those resources may create additional solar or battery opportunities around the project.
The objective is to create a clearer view of the project's energy path:
Initial Capacity
↓
Interconnection
↓
Utility Infrastructure
↓
First Energization
↓
Additional Generation and Storage
↓
Expansion Capacity
↓
Full Campus Build-Out
That roadmap can help the development team identify which energy decisions need attention first.
AI data centers can move from concept to construction quickly.
Grid infrastructure often cannot.
That mismatch is one of the most important development risks facing large data center projects.
A utility may need time to:
For developers, that means the grid schedule can become the project schedule.
A strong grid strategy should identify which pieces of infrastructure control energization.
That may include:
Once those dependencies are understood, the project team can begin planning around them.
Some campuses may be able to receive an initial block of power before the full planned load is available.
That can create a phased energization strategy.
For example:
Phase 1
Use available utility capacity to energize the first portion of the campus.
Phase 2
Add capacity as new utility infrastructure comes online.
Phase 3
Expand further as additional substations, transmission capacity, or generation becomes available.
The correct sequence depends on the project and utility.
But phased energization can help align data center construction with the actual pace of grid development.
A power strategy can also be delayed by physical equipment.
Transformers, switchgear,r and other specialized electrical components may have substantial procurement timelines.
That means the project should understand not only when the utility expects to complete its work, but also whether critical equipment can be sourced in time.
Important grid milestones may include:
These should be tracked alongside permitting, construction, and tenant or compute deployment schedules.
There may be several relevant dates:
A useful grid strategy creates a roadmap for each stage rather than relying on one final energization date.
When a utility identifies limited capacity or a long upgrade timeline, it is easy to view the issue as a binary outcome:
Either the grid can serve the project or it cannot.
In practice, there may be additional questions worth asking.
If full build-out requires more capacity than the system can provide immediately, a phased development plan may allow the project to begin with a smaller initial load.
That could give the utility additional time to complete infrastructure improvements.
Battery energy storage may support portions of the project strategy by helping manage demand or reshape when electricity is drawn from the grid.
Potential applications may include:
Battery storage does not automatically create new grid capacity.
Its value depends on how the system is designed, operated, and treated by the utility.
But it may create additional flexibility worth evaluating.
Solar generation may help offset portions of daytime electricity demand.
Potential locations can include:
Solar is not a substitute for the continuous power required by a large AI data center.
But where local generation reduces portions of demand or supports a broader energy strategy, it may become part of the solution.
The data center property is not always the only place to look.
The surrounding market may contain:
These resources may not directly eliminate the need for grid upgrades.
They can, however, add generation and storage throughout the local energy system and create more options around the broader project.
Some projects may benefit from combining:
The goal is not to create complexity for its own sake.
It is to reduce unnecessary dependence on a single assumption about how and when power will become available.
If the traditional power path works, that may remain the best solution.
If it does not, the project team should understand what other resources are available before concluding that the site cannot move forward.
The strongest strategy identifies grid constraints early enough to preserve alternatives.
AI data centers rarely remain static.
A campus may begin with one building, one phase, or one block of electrical capacity, then expand as additional compute infrastructure comes online.
That means the grid strategy should not be designed only around the first energization milestone.
It should also answer:
What happens when the project needs more power?
Developers should understand the expected electrical requirements at full build-out.
That may include:
A site that can support the first phase but cannot realistically support later expansion may create a long-term development problem.
Future electrical capacity may require additional physical infrastructure.
That can include:
Those requirements should be considered early in site planning.
Land that appears unnecessary during the first phase may become critical later.
The development team should understand what additional utility capacity might require.
Questions may include:
The answers can affect how quickly future phases can move forward.
A broader energy strategy may also expand over time.
Future phases could include:
The energy roadmap should evolve alongside the data center.
One of the risks of focusing only on initial energization is that the project may encounter the same grid constraint again during expansion.
A stronger strategy considers the full development path from the beginning.
The goal is not simply to connect the first building.
It is to create a credible grid path for the entire campus.
A data center may be located on a single parcel, but the electrical infrastructure serving it can extend miles beyond the site.
That makes substations, transmission, and distribution infrastructure central to development planning.
Substations convert and route electricity between different parts of the grid.
For a large AI data center, available substation capacity can be one of the most important site-selection factors.
A project may need to determine:
A nearby substation does not necessarily mean sufficient capacity exists.
Some projects may require a dedicated or expanded substation.
That can introduce additional considerations involving:
A new substation can become one of the largest infrastructure components of the project.
Very large loads may also depend on the capacity of the higher-voltage transmission system.
If the local transmission network cannot support the additional demand, larger system upgrades may be required.
Those upgrades can be more complex and may affect timelines well beyond the immediate data center site.
Not every project will connect directly at transmission voltage.
Distribution-level infrastructure can also become important depending on the size and structure of the facility.
That may involve:
The correct architecture depends on the facility’s load, utility standards, and reliability requirements.
Data centers have unusually high reliability expectations.
That can influence how many electrical feeds, substations, or other redundant systems are required.
The project may need to evaluate:
Reliability requirements should be integrated into the grid strategy from the beginning.
The data center site plan may need space for:
Waiting until late design to account for these assets can create avoidable conflicts with buildings, roads, drainage, landscaping,g or future expansion.
For AI infrastructure, grid assets are not secondary support equipment.
They are fundamental parts of the development.
The sooner substations, transmission requirements, and electrical corridors are understood, the more effectively the overall site can be designed around them.
A data center can secure an attractive power contract and still have a development problem if the electricity cannot physically reach the site.
Likewise, a site may have strong grid infrastructure but an incomplete long-term power procurement strategy.
That is why grid and procurement planning should be coordinated.
The grid strategy determines how electricity reaches the facility.
That may involve:
The procurement strategy determines where the electricity comes from and under what commercial terms.
That may involve:
Both sides have to work.
A renewable energy agreement can support long-term procurement and sustainability goals.
But signing a PPA does not automatically increase the physical capacity of the local grid serving the data center.
The project still needs:
This distinction is important.
The commercial power strategy and physical grid strategy should be evaluated together.
Distributed energy resources introduce their own grid questions.
Solar and battery projects may require:
A promising solar or storage opportunity is not complete until the interconnection path is understood.
A diversified power portfolio may make the overall strategy more resilient.
Depending on the project, that portfolio could include:
The grid strategy determines how those resources interact with the physical electricity system.
Interconnection, procurement, solar, storage, and expansion should not be planned as unrelated projects.
They are parts of one energy system.
The strongest strategy aligns the commercial power plan with the physical path that delivers electricity to the data center.
For a large AI data center, grid strategy should begin during site evaluation or early development.
Waiting until design and permitting are already underway can leave the project with fewer options if the local electricity system cannot support the required load on the expected timeline.
The earlier the grid path is understood, the more effectively the development team can plan around it.
A potential site should be evaluated for more than proximity to transmission lines or substations.
Important questions may include:
A site that appears ideal from a real estate perspective can become difficult if its grid path is weak.
Electrical infrastructure requires physical space.
The site plan may need to accommodate:
If these requirements are discovered too late, they can conflict with buildings, roads, stormwater systems, setbacks, or future expansion areas.
Utility discussions can establish assumptions that shape the project for years.
Before major commitments are finalized, developers should understand:
The goal is not to delay utility engagement.
It is to make sure the development team understands the implications of the proposed grid path before it becomes difficult to change.
Existing data centers should also revisit grid strategy before major expansion.
A campus that originally had sufficient power may face new constraints when:
The grid strategy should evolve with the facility.
The project may ultimately follow the conventional utility path.
That may be the best solution.
But understanding the grid early gives the development team time to evaluate alternatives, phasing, and additional resources before the power schedule begins controlling the project.
A data center interconnection strategy naturally focuses on the point where the facility connects to the grid.
The broader energy opportunity can extend much farther.
Solar Harmonics also looks at the surrounding market to identify resources that may support a more flexible local energy strategy.
Potential opportunities may include:
These resources do not automatically eliminate the need for utility upgrades.
Their value is different.
They may add local generation, storage,e and flexibility around the same market where the data center is creating major new demand.
Nearby warehouses and industrial facilities may have:
A group of individual projects may collectively create a meaningful distributed energy portfolio.
Public agencies may also control suitable facilities or land.
Potential opportunities can include:
Where practical, these projects can connect data center energy planning with broader local energy goals.
Standalone or distributed storage may also be valuable in some markets.
Potential value depends on:
The role of each battery project has to be evaluated individually.
This distinction is important.
Distributed solar and batteries should not be presented as a guaranteed substitute for transmission, substations, or utility infrastructure.
A large data center may still require substantial grid investment.
The opportunity is to determine whether additional energy resources can strengthen the broader system surrounding that investment.
A data center creates a reason to invest in major electricity infrastructure.
The Solar Harmonics Framework asks whether that same development can also help accelerate additional energy investment throughout the surrounding community.
Instead of viewing the project only as:
New Load → More Grid Infrastructure
the broader strategy can explore:
New Load → Grid Investment + Local Generation + Storage + Community Energy Opportunity
That does not change the physics of the grid.
It changes the scope of the energy strategy.
Grid access can affect nearly every major element of an AI data center development.
It can influence:
That makes grid and interconnection planning one of the most important early development workstreams.
Solar Harmonics helps developers evaluate the energy environment surrounding a proposed or expanding California data center.
An initial review can examine:
The strongest grid strategy is not simply the one that identifies a connection point.
It is the one that gives the development team a realistic understanding of:
That is the information developers need before grid constraints begin driving project decisions.
If you are planning, developing, or expanding an AI data center in California, Solar Harmonics can help evaluate the grid environment, interconnection path, and additional energy opportunities surrounding the project.
Data center interconnection is the process of connecting a data center's electrical load to the utility or power-provider system that will serve it.
For a large AI data center, interconnection can involve far more than requesting standard electric service.
The process may require:
The exact process varies by utility, municipal power provider, project size and service voltage.
For large AI projects, interconnection should be treated as a development workstream with its own schedule, costs and risks.
As early as possible.
Ideally, grid and interconnection planning should begin during site selection.
A site may appear attractive because of:
But if the power path is weak, those advantages may not be enough.
Early grid analysis can help determine:
The later these issues are discovered, the fewer alternatives a developer may have.
There is no single standard timeline.
The schedule depends on:
A project that can use existing infrastructure may move more quickly than one that requires a new substation or major transmission upgrades.
For planning purposes, developers should avoid focusing only on the final energization date.
It is often more useful to understand several milestones:
That provides a more realistic view of the power path.
Time to power is the period between project development and the point when sufficient electricity becomes available to operate the data center.
For AI infrastructure, this can become one of the most important site-selection and development factors.
A data center may be capable of being designed and constructed faster than the grid infrastructure required to serve it.
Time to power may therefore depend on:
The critical question is not only whether power can eventually be provided.
It is whether the required capacity can become available on the project's development timeline.
A utility interconnection study evaluates how a proposed project may affect the electric system and what infrastructure may be required to serve it.
Depending on the project, the study may examine:
The study may identify upgrades that are necessary before the project can receive its requested load.
For a large data center, the results can influence:
Transmission generally refers to the high-voltage system used to move large amounts of electricity over longer distances.
Distribution refers to the lower-voltage network that delivers electricity closer to end users.
Large AI data centers may interact with both.
Depending on project size and utility design, a facility may require:
The specific architecture varies significantly by project.
Not always, but some large projects may require a dedicated or significantly expanded substation.
The need depends on:
A dedicated substation can provide substantial capacity, but it also introduces:
The substation strategy should be understood early in the development process.
There is no universal acreage requirement.
The footprint depends on:
The key development issue is to reserve enough space early.
If substation and electrical-yard requirements are added after the site plan is largely complete, they can conflict with:
Electrical infrastructure should therefore be part of the site-planning process from the beginning.
Data centers have unusually high reliability requirements.
For many facilities, a brief interruption in electricity can have significant operational consequences.
That can lead developers to evaluate:
Redundancy requirements can affect both the grid connection and the amount of infrastructure needed on the site.
They should be considered when evaluating capacity rather than added after a basic service design is complete.
Yes.
Phased energization can be useful when the full electrical load is not available immediately.
A project might receive:
Initial Capacity
Enough power to open the first phase.
Additional Capacity
More electricity as utility upgrades are completed.
Full Build-Out Capacity
The ultimate load after major infrastructure is in place.
This can allow data center construction and grid development to proceed on different but coordinated schedules.
A phased strategy only works if later capacity has a credible path.
The developer should understand what infrastructure, agreements, and construction are required for each phase.
Yes.
The initial interconnection should be evaluated in the context of the full expected campus load.
Questions may include:
A site that supports the first phase but has no practical expansion pathway can become constrained later.
That depends on the utility and project structure.
Developers should not assume that future capacity will automatically remain available simply because the first phase is approved.
The project may need to understand:
Future capacity should be discussed explicitly with the serving power provider.
Major electrical equipment can become part of the critical path.
Examples include:
If required equipment has a long procurement timeline, energization may be delayed even after engineering and utility approvals are complete.
That is why equipment planning should be coordinated with the interconnection schedule.
Power procurement determines how the data center plans to source electricity commercially.
Grid interconnection determines how electricity physically reaches the site.
A project needs both.
For example:
A data center could sign a renewable PPA but still need a utility substation and transmission capacity.
A facility could have access to strong grid infrastructure but still need a long-term procurement strategy for cost, renewable energy, and expansion.
Solar Harmonics treats the two as connected parts of the same power strategy.
A credible path to power is more than a statement that electricity may eventually be available.
It should provide a reasonable understanding of:
For a phased project, it should also explain how the campus moves from the first block of power to the ultimate load.
The more clearly those steps are understood, the more confidently the development team can plan around them.
No.
Physical proximity to a substation does not guarantee usable capacity.
A nearby substation may already have:
The project may still require:
A substation can be close geographically and still be unable to support the required load without significant investment.
No.
Transmission lines indicate that major electrical infrastructure is nearby, but that does not mean the system has available capacity for a large new load.
Transmission capability depends on broader system conditions.
A project may still face limitations involving:
Transmission proximity is useful, but it should not be treated as proof that the site can support a large AI data center.
The required infrastructure depends on project size, location, and utility standards.
Potential infrastructure may include:
Very large projects may require dedicated infrastructure that becomes a major part of the overall development.
That depends on the utility, tariff, project structure,e and the type of infrastructure being built.
Potential costs may include:
Some infrastructure may primarily serve the data center.
Other improvements may provide broader grid benefits.
Cost allocation can be a major factor in site economics, which is why developers should understand these requirements before making major commitments.
Potentially, but battery storage should not be treated as automatic new grid capacity.
Battery systems can create flexibility by:
Whether storage can help a specific interconnection depends on:
The utility may still require the grid to be designed around the facility's full expected demand.
Battery value should therefore be evaluated project by project.
Solar can reduce portions of electricity demand while it is generating, but it does not automatically eliminate the need for grid capacity.
A large AI data center can operate continuously, while solar generation changes throughout the day.
Solar may help:
The project may still require substantial utility capacity for:
Solar should be incorporated into grid strategy based on actual generation and load profiles.
They may support the broader energy strategy.
Distributed energy resources can include:
These resources may add renewable generation and storage around the same market where the data center is creating new electricity demand.
They do not automatically remove the need for transmission, substations, or utility upgrades.
Their value is in creating additional energy resources and potentially more flexibility around the broader project.
Not by itself.
A data center can contract for large amounts of renewable energy while still facing a physical grid constraint at its site.
A PPA or renewable energy contract addresses where energy is procured commercially.
Grid capacity addresses whether sufficient electricity can physically be delivered to the facility.
Both need to work.
That is why data center power procurement and interconnection strategy should be planned together.
Generation capacity refers to the ability to produce electricity.
Grid capacity refers to the ability of transmission, substations, and distribution infrastructure to move that electricity where it is needed.
A region may have significant generation and still have local grid constraints.
For a data center developer, the relevant issue is whether enough usable power can reach the specific site at the required time.
The project may need to evaluate several possibilities.
Potential options can include:
The correct response depends on the size and nature of the constraint.
The important step is to identify the problem early enough that alternatives remain available.
No.
Solar Harmonics does not replace the serving utility, electrical engineer, transmission planner, or other specialized technical professionals responsible for formal grid studies and engineering.
Our role is strategic.
We help developers evaluate the broader energy environment around the project and identify the grid questions, infrastructure issues, and additional energy opportunities that should be understood early.
That can include:
Formal engineering and utility determinations remain with the appropriate technical and utility organizations.
The Solar Harmonics Framework expands the grid conversation beyond the data center connection itself.
The traditional model can be viewed as:
Data Center Demand → Utility Infrastructure → Grid Power
The Solar Harmonics Framework explores a broader model:
Data Center Demand
Utility or Municipal Power
Solar Generation
Battery Storage
Distributed Energy Resources
Local Property Participation
Community Energy Investment
The data center may still require substantial utility infrastructure.
The framework asks whether that major new load can also become a catalyst for additional generation and storage throughout the surrounding market.
Useful information can include:
Not every item needs to be available before the first conversation.
An initial review can help determine which unknowns are most important to investigate.
The goal is not simply to obtain a utility connection.
It is to create a realistic power path that supports the data center from initial development through full build-out.
That means understanding:
The strongest grid strategy gives the development team enough information to make better decisions before electricity becomes the constraint controlling the project.
Key Points San Jose has spent decades at the center of the digital economy. Now the capital of Silicon Valley is confronting the physical infrastructure
Key Points Pittsburg, California, is emerging as an important test case for one of the biggest questions facing the artificial intelligence economy: How can communities
Key Points Hayward, California, is becoming part of a much larger transformation taking place across the San Francisco Bay Area. Artificial intelligence is driving demand
Key Points The San Francisco Bay Area helped create the technology behind today’s artificial-intelligence revolution. Now the region faces a different challenge: How will we
Key Points Introduction A 500 MW data center is not an ordinary commercial development. When a major AI infrastructure developer approaches a city, the proposal
Key Points The race to build artificial intelligence infrastructure is increasingly becoming a race for electricity. Land can be acquired. Buildings can be constructed. GPUs
