Energy & Utilities.
From traditional hydrocarbons to nuclear renaissance to the AI-driven power crisis — we analyze energy through the lens of physical infrastructure constraints, regulatory frameworks, and the unprecedented demand inflection from data center electrification.
Our Thesis
Energy is experiencing a demand inflection unlike anything in the past fifty years. After two decades of flat-to-declining U.S. electricity demand, the convergence of AI data center buildout, electric vehicle adoption, manufacturing reshoring, and electrification of buildings is creating a structural step-change in power consumption that the existing grid was not designed to handle. The Energy Information Administration projects U.S. electricity demand will grow 15-20% by 2030 — a rate not seen since the 1970s. This is not a forecast based on speculative adoption curves; it is driven by capital commitments already made and construction already underway.
The investment implications are profound. Every electron consumed by an AI data center must be generated, transmitted, and distributed through physical infrastructure. Utilities must build generation capacity (natural gas, nuclear, renewables), expand transmission networks, and upgrade distribution systems. Power equipment manufacturers must produce transformers (lead times now exceed 2-3 years for large power transformers), switchgear, circuit breakers, and control systems at volumes that exceed current production capacity. This creates a multi-year investment cycle across the entire power value chain.
We are particularly focused on the nuclear renaissance. Nuclear power provides the only scalable source of carbon-free baseload electricity — exactly what AI data centers require (reliable, 24/7, high-capacity-factor power). Microsoft has signed a 20-year PPA to restart Three Mile Island Unit 1. Amazon has invested in nuclear-powered data centers. Google has partnered with Kairos Power on advanced reactors. Small modular reactors (SMRs) from NuScale, GE Hitachi, and others promise to deliver factory-built nuclear plants at smaller scale and lower capital cost. The nuclear supply chain — from uranium mining to fuel fabrication to reactor construction to waste management — represents a generational investment opportunity that the market is only beginning to recognize.
Research Framework
Energy Transition Modeling
- •Renewable energy LCOE trends: solar, onshore wind, offshore wind cost curves by region
- •Intermittency analysis: capacity factor vs. nameplate, curtailment, storage requirements
- •Baseload requirement modeling: what percentage of grid load requires firm, 24/7 power
- •Natural gas as bridge fuel: LNG export economics, pipeline capacity, associated gas production
- •Grid-scale storage: lithium-ion, iron-air, compressed air, pumped hydro — cost and duration analysis
Nuclear Economics
- •Large reactor economics: construction cost overruns (Vogtle lessons), regulatory timeline risk
- •Small modular reactors (SMR): NuScale, GE-Hitachi BWRX-300, X-energy, Kairos — design maturity comparison
- •Gen IV advanced reactors: molten salt, high-temperature gas, sodium fast reactors
- •Nuclear fuel cycle: uranium mining (ISR, conventional), enrichment (SWU pricing), fuel fabrication
- •Plant life extensions: NRC license renewal economics, 80-year operating life potential
Utility Regulatory Frameworks
- •Rate case analysis: allowed ROE, rate base growth, capital recovery mechanisms
- •Power purchase agreement (PPA) structures: corporate, utility-scale, behind-the-meter
- •Grid interconnection queue analysis — 2,600+ GW backlog, reform proposals, timeline implications
- •Renewable portfolio standards (RPS) and clean energy mandates by state
- •Transmission planning: FERC Order 1920, regional planning, interregional transfer capacity
Power Equipment Demand
- •Large power transformer shortage: 2-3 year lead times, limited domestic manufacturing capacity
- •Switchgear and circuit breaker demand: data center, industrial, utility segments
- •Medium-voltage electrical equipment: the data center power chain from utility to rack
- •Grid modernization: smart grid, DERMS, advanced metering, distribution automation
- •Domestic manufacturing incentives: IRA production tax credits, Defense Production Act invocations
How We Analyze the Data Center Power Chain
Step 1: Demand Quantification
We build bottom-up demand models starting with hyperscaler capex commitments and data center construction permits. A typical 100MW AI data center campus draws power equivalent to approximately 80,000 homes. When Microsoft commits to building 50+ GW of data center capacity by 2030, that represents roughly 40 million home-equivalents of incremental power demand from a single company. We aggregate demand across all hyperscalers, colocation providers, and enterprise data centers, overlay it on existing utility capacity maps, and identify the specific regions and utilities facing the largest demand-supply gaps. Northern Virginia (Dominion Energy territory), Dallas-Fort Worth (Oncor), central Ohio (AEP), and Phoenix (APS/SRP) are the primary pressure points today, but new clusters are emerging in Mississippi, Indiana, and the Carolinas as primary markets saturate.
Step 2: Generation Source Analysis
Data centers require high-capacity-factor, reliable power — intermittent renewable energy alone cannot serve a facility that runs 24/7/365 at near-full load. We model the generation mix needed to serve AI data center demand: natural gas combined-cycle plants (high capacity factor, moderate carbon intensity, 3-4 year build time), nuclear (highest capacity factor, zero carbon, 7-10 year build time for new plants or 2-3 years for restarts/uprates), and renewables-plus-storage (zero marginal cost, but requires 3-4x nameplate overbuild plus multi-hour storage to achieve firm capacity equivalent). Each generation source has a different cost structure, construction timeline, permitting pathway, and carbon profile. Hyperscalers with clean energy commitments face a particularly acute challenge: they need firm power now, but their preferred sources (nuclear, renewables-plus-storage) take the longest to deploy.
Step 3: Transmission & Interconnection
Even where generation capacity exists, transmission infrastructure must deliver it to the data center. The U.S. grid interconnection queue contains over 2,600 GW of proposed projects — mostly renewables and storage — but the average project spends 5 years in the queue and only 14% of projects that enter the queue ultimately reach commercial operation. We track queue positions, completion rates by region (PJM, ERCOT, CAISO, MISO), and transmission construction timelines. FERC Order 1920 on regional transmission planning could accelerate buildout, but implementation takes years. In the interim, data centers are increasingly co-locating with generation sources (building gas plants or nuclear behind the meter) to bypass grid constraints entirely — a trend that raises regulatory and grid reliability questions but solves the immediate power access problem.
Step 4: Equipment Supply Chain Assessment
The physical equipment that constitutes the power chain — transformers, switchgear, circuit breakers, bus duct, automatic transfer switches, UPS systems, and power distribution units — is manufactured by a concentrated group of companies with limited capacity to surge production. Large power transformers (LPTs) are the most constrained: only a handful of factories worldwide produce them, lead times now exceed 2-3 years, and domestic U.S. production capacity covers only a fraction of demand. Powell Industries manufactures custom-engineered switchgear and is seeing unprecedented demand from data center, utility, and industrial customers. Vertiv produces thermal management and power systems purpose-built for high-density computing environments. We assess order backlogs, capacity expansion plans, and pricing power for each segment of the power equipment chain.
Step 5: Valuation Across the Energy Spectrum
Energy companies span a wide valuation framework range: regulated utilities on P/E and dividend yield, independent power producers on EV/EBITDA and contracted cash flow, oil & gas on EV/reserves and FCF yield, power equipment on revenue growth and margin expansion, and nuclear developers on option value and milestone de-risking. We apply the appropriate framework to each subsector while maintaining a unifying thesis: the structural increase in electricity demand creates a multi-year tailwind for companies across the power value chain. The key is identifying which companies have genuine capacity constraints (pricing power), which benefit from regulatory tailwinds (rate base growth, production tax credits), and which are commodity participants that will see revenue growth but not margin expansion. We are willing to pay premium multiples for companies at genuine supply chain chokepoints and seek value in overlooked infrastructure plays trading at industrial multiples despite utility-grade demand visibility.
Coverage Universe
- Cameco (CCJ)
- Constellation Energy (CEG)
- Vistra (VST)
- NANO Nuclear (NNE)
- Oklo (OKLO)
- Bloom Energy (BE)
- ExxonMobil (XOM)
- Chevron (CVX)
- ConocoPhillips (COP)
- Pioneer Natural Resources (PXD)
- Diamondback Energy (FANG)
- Southern Company (SO)
- NextEra Energy (NEE)
- Duke Energy (DUK)
- Dominion Energy (D)
- AES Corporation (AES)
- Powell Industries (POWL)
- Vertiv Holdings (VRT)
- Eaton Corporation (ETN)
- Hubbell (HUBB)
- Schneider Electric (SBGSF)
- Enphase Energy (ENPH)
- First Solar (FSLR)
- Fluence Energy (FLNC)
- Volta Green (VG)
- IMSR (IMSR)
- XLE (Energy Select)
- XLU (Utilities Select)
- URNM (Uranium Miners)
- ICLN (Clean Energy)
- TAN (Solar)
Current Themes
AI Data Center Power Crisis
The single most disruptive force in U.S. energy markets is AI-driven data center power demand. Total U.S. data center power consumption is projected to grow from 17 GW to 35+ GW by 2030, with some estimates exceeding 50 GW. Utilities in key data center markets are filing emergency capacity expansion requests with regulators. Transformer lead times have stretched to 3+ years. Hyperscalers are signing 10-20 year power purchase agreements at premium prices to secure capacity. This demand shock benefits every company in the power value chain — generation, transmission, distribution, and equipment manufacturing — for the remainder of the decade.
Nuclear Renaissance
Nuclear power has been reclassified from a liability to a strategic asset. Microsoft's restart of Three Mile Island, Amazon's nuclear-powered data center investments, and Google's partnership with Kairos Power signal that hyperscalers view nuclear as essential infrastructure. Constellation Energy and Vistra are seeing existing nuclear fleet valuations reprice dramatically. NuScale, GE-Hitachi, X-energy, and Oklo are advancing SMR designs through NRC review. The nuclear fuel supply chain — uranium miners (Cameco, Kazatomprom), enrichment (Urenco, Orano), and fuel fabrication — faces structural undersupply as secondary sources deplete and primary demand accelerates.
LNG Export Expansion
U.S. natural gas production continues to grow, driven by associated gas from Permian Basin oil production and efficiency gains in dry gas basins. LNG export capacity is expanding with new terminals on the Gulf Coast and potentially the West Coast. Global LNG demand is being pulled by Asian industrialization, European energy security (replacing Russian pipeline gas), and the need for gas-fired generation to back up intermittent renewables worldwide. Companies positioned in natural gas production, pipeline infrastructure, and LNG liquefaction capture a structural demand trend that extends well beyond the U.S. market.
Grid Modernization Imperative
The U.S. electrical grid was designed for one-directional power flow from centralized generation to passive consumers. The combination of distributed generation (rooftop solar), bidirectional flows (battery storage, EV V2G), demand response, and massive new point-loads (data centers, EV charging) requires fundamental grid modernization. Utilities are investing in smart grid technology, advanced metering infrastructure, distribution automation, and grid-edge intelligence. Companies providing grid modernization equipment and software are positioned for a multi-decade investment cycle that is just beginning.
Interested in our energy research?
All inquiries are treated with strict confidentiality.
Investor Inquiries