Energy Model
How will AI be powered? Datacenter electricity demand is already transforming electricity as a commodity, power grids as critical infrastructure, and having ripple effects across the entire energy supply chain. This model is built for investors, IPPs, hyperscalers, developers, and utilities.
- Scope
- United States
- Generating units
- 25,000+
- Tracking
- 30+ OEMs; 60+ BAs
- Coverage
- 2002–2040
Where power grids fall short.
The SemiAnalysis Energy Model takes the datacenter model and translates it into power infrastructure demand. A national inventory of electric generators, storage, and onsite energy solutions is integrated with datacenter IT capacity projections and the integrated resource plans of electric utilities, RTOs, and ISOs, producing a granular view of the supply-demand shortfall in every major US electricity market.
The model then integrates this view with supply chain analysis for generation units that are crucial to behind-the-meter solutions. Clients see the power bottlenecks in datacenter demand growth, and where those constraints can be overcome. The fastest-growing solution is behind-the-meter generation. The Energy Model tracks these generating units order by order; including every disclosed onsite gas or fuel cell project at a US datacenter site, the equipment behind it, and the OEM manufacturing capacity that limits how fast it can scale.
Who needs this model and why:
- Utilities & IPPs
- To understand what datacenters mean for their markets; so that they can mitigate capacity shortfall risks, seize revenue opportunities, and keep up with the demand-driven transformation of their sector.
- Hyperscalers & datacenter developers
- To know where to go and whether others have gone; as well as keep apace of regulatory developments that will impact timelines and bottom-lines.
- Equipment manufacturers
- To track competitors, identify market demand and key customer opportunities, and develop a view on saturation risk.
- Public & private market investors
- To take positions across utilities, IPPs, and power equipment names integrated with SemiAnalysis’ expertise across datacenter demand, AI-economics, and semiconductor.
Power-to-Compute Crunch
AI demand sets compute demand; compute demand sets power demand; and power demand is now hitting an ill prepared power grid. SemiAnalysis specialises in each link of the chain.
Demand
Datacenter load built from the SemiAnalysis datacenter pipeline, facility by facility; adjusted for PUE and matched to locational reserve margins. This position is also contrasted against utility positions.
Load forecasts
Datacenter load by market and service territory, adjusted for PUE and reserves, with the named facilities behind every megawatt.
Forecasts on forecasts
The SemiAnalysis demand view set against official utility and RTO projections, showing who is under-calling the load and whose stated pipeline exceeds what can be served.
Regulatory risks
Regulatory frameworks and updates in markets like ERCOT and PJM, where changes carry implications for datacenter large load projects
Supply
A national inventory of what can actually generate, store, and deliver power, translated into the firm capacity available to meet new load:
Generator inventory
Existing and planned generation, storage, and onsite energy solutions across the United States
Firm capacity
Nameplate translated to firm capacity, so intermittent resources count for what they can actually deliver
Projects pipeline
New generation and transmission projects tracked from announcement through energization
The Shortfall
Supply-demand shortfall analysis for datacenter load growth across US electricity markets. Each year’s new datacenter demand is set against each year’s new accredited supply, market by market. Focusing on which markets run short first, when and where the supply-demand shortfall hits, and where supply or transmission constraints delay load growth.
Solutions
The market finds a way to meet demand. The largest solution is behind-the-meter power, which the model tracks as a market of its own: every disclosed onsite generation and storage order at a US datacenter site is logged with its OEM, megawatts, and delivery timeline, and set against each manufacturer’s production ceiling and sold-out horizon. Other solutions include: storage, flexible and non-firm connections, and the new generation the shortfall itself is pulling into the queue. New datacenters are increasingly built as energy parks, hybridizing a grid connection with generation on site.
Where the grid runs short.
Market by market, the model sets projected datacenter load against the firm capacity available to serve it. The overshoot is the gap, and the gap is where the decisions live.
Bars are illustrative. In the model, each market’s gap is quantified utility by utility, with the generation and transmission pipeline set against it and the solutions most likely to bridge it.
How the model is built.
Demand and supply are built independently, then reconciled.
-
Forecast the load
Datacenter IT capacity projections translate into utility-level load, adjusted for PUE and reserve margins, and compared against each utility’s own forecast.
-
Inventory the supply
Existing and planned generation, storage, and onsite energy are catalogued nationally and translated to firm capacity, with grid projects tracked to energization and behind-the-meter orders tracked to delivery.
-
Reconcile and flag
Load meets capacity market by market: the model flags where the grid falls short, which constraints delay growth, and what may bridge each gap.
Research that ships with the model.
Model subscribers receive the update notes, webinars, and analysis published against each release.
Common questions.
Anything not covered here, ask the team directly through the form below.
What does the Energy Model include?
Top-down utility-level datacenter load forecasts adjusted for PUE and reserves, a bottom-up national inventory of generation, storage, and onsite energy translated to firm capacity, tracking of new generation and transmission projects, order-by-order tracking of behind-the-meter generation and storage at US datacenter sites, regulatory-framework coverage of markets like ERCOT and PJM, and supply-demand gap analysis across major US electricity markets.
How is the model built?
A national inventory of electric generators, storage, and onsite energy solutions is integrated with SemiAnalysis datacenter IT capacity projections and the integrated resource plans of electric utilities, RTOs, and ISOs. Demand is then compared against each utility’s own forecasts, so the gaps are visible before they show up in filings.
How is the model delivered?
As an Excel workbook with dashboard access, including one year of quarterly updates, an onboarding call with the team to walk through the model and methodologies, and ad-hoc calls for questions that come up in use.
Is it part of the SemiAnalysis newsletter subscription?
No. Industry models are separate institutional offerings and are not included with the annual newsletter membership.
Can it show which utilities run short of power?
Yes. The model provides a forward-looking view of which utilities face the most severe supply-demand gaps, where and how new generation capacity comes online, and where power supply and transmission constraints may delay load growth, along with the solutions that may bridge each gap.
Models that pair with this one.
Pairing depends on the decision in front of you. The load this model sets against the grid is created by the facilities and hardware the other models cover.
Datacenter Industry Model
Demand sourceThe site-level facility pipeline creating the load this model sets against the grid.
View model
Accelerator & HBM Model
Demand forecastThe accelerator shipments whose power draw starts the chain this model follows to the grid.
View model
AI Cloud TCO Model
Cost modelWhere power prices land in the all-in cost of a GPU-hour, cluster by cluster.
View modelGet the Energy Model.
Start with the sales team. They come back with scoping, licensing, and pricing for your mandate.
- Scoped to your use case
- Onboarding and ad-hoc analyst calls included
- Custom research engagements available
