ENERGY ECONOMICS

Energy economics is the study of how commercial energy is generated, transmitted through utility grids, priced in competitive wholesale markets, and allocated across industrial infrastructure, directly governing capital allocation and global computational scale.

Baseload vs. Intermittent Energy: Generation Profiles Compared

The economic viability of continuous, 24/7/365 industrial loads—such as semiconductor foundries and hyperscale AI compute facilities—is determined by the physical characteristics of regional generation assets:

Generation Asset	Reliability Profile	LCOE ($/MWh)	Interconnection & Build Timeline	Carbon Footprint ($gCO_2e/kWh$)	Primary Economic Risk
Nuclear Fission / SMR	Firm Baseload (>90% capacity factor)	$90 – $160	6 – 10+ years	< 12	Regulatory delays, high initial CapEx
Geothermal Energy	Firm Baseload (>90% capacity factor)	$60 – $100	4 – 6 years	< 40	Subsurface exploration risk
Natural Gas (CCGT)	Firm Dispatchable (>85% capacity factor)	$45 – $75	3 – 5 years	~400	Fuel price volatility, carbon taxation
Utility Solar + BESS	Intermittent (25% capacity factor)	$35 – $60	2 – 4 years	< 25	Battery degradation, weather seasonality
Hydroelectric	Firm Dispatchable (>60% capacity factor)	$40 – $70	8 – 15 years	< 20	Drought vulnerability, geographic scarcity

The Mathematics of Datacenter Power Economics

Operating costs for large-scale digital infrastructure are directly governed by power consumption, facility efficiency, and contracted electricity rates.

Total annual facility energy consumption (E_annual, in Megawatt-hours) is modeled as:

E_annual = P_nameplate × 8,760 × C_f × PUE

Where:

• P_nameplate is the rated IT power capacity in Megawatts (MW).
• 8,760 represents total hours in a non-leap calendar year.
• C_f is the operational capacity factor (typically 0.85 to 0.95 for continuous inference and training).
• PUE is Power Usage Effectiveness (Total Facility Power / IT Equipment Power).

Financial Impact of Efficiency Gains

Consider a 100MW campus operating at C_f = 0.90 with wholesale power contracted at $60/MWh:

• Standard Facility (PUE = 1.40): Consumes 1,103,760 MWh/yr, incurring an annual electricity bill of $66.23 million.
• Liquid-Cooled Facility (PUE = 1.15): Consumes 906,660 MWh/yr, incurring an annual electricity bill of $54.40 million.

Compressing PUE by optimizing facility thermodynamics and liquid cooling yields $11.83 million in annual direct OpEx savings per 100MW, creating an unassailable margin advantage for disciplined operators.

The Interconnection Queue Bottleneck and Power Moats

The primary competitive friction in modern infrastructure is not hardware procurement, but utility interconnection delays.

In major transmission jurisdictions (such as PJM, ERCOT, and CAISO in the United States), applications to connect large industrial loads face backlog queues spanning 5 to 7 years due to inadequate high-voltage transmission lines and substation transformer shortages.

Consequently, industrial operators that secured long-term Power Purchase Agreements (PPAs) and approved grid interconnections prior to the generative AI buildout hold a structural economic moat. These "energized shovel-ready" land sites command valuations 3x to 5x higher than un-interconnected commercial parcels.

Strategic Responses: Behind-the-Meter and Nuclear PPAs

To bypass utility interconnections and maintain 100% clean baseload operational targets, hyperscalers are rewriting enterprise capital allocation playbooks:

1. Behind-the-Meter Colocation: Building computational facilities directly adjacent to nuclear power plants or dedicated natural gas generation, drawing electricity before it touches the regulated public grid.
2. Small Modular Reactor (SMR) Anchor Offtake: Committing billions in forward capital guarantees to finance factory-fabricated 50MW–300MW nuclear reactors with anticipated deployment windows in the early 2030s.
3. Firm Geothermal Contracts: Partnering with next-generation enhanced geothermal developers utilizing directional drilling to unlock round-the-clock renewable generation regardless of surface topography.

Through this structural evolution, electricity procurement has transformed from an administrative utility expense into the decisive strategic bottleneck dictating the velocity and geographical distribution of global computing power.

Part of the knowledge graph at The Best Blog Ever — reference definitions for ideas that matter.