Thematic Exposure -- 5/10

Talen is exposed to one of the best secular themes in the market — AI/data-center-driven power demand and a tightening PJM grid — but it competes in a fragmented merchant-power market where it holds only ~7% of PJM capacity (~13 GW of ~182 GW) and sets no prices. It is a price-taker in both the energy (LMP) and capacity (RPM) auctions, trailing Dominion, Constellation, and Vistra. The theme is A-grade; the competitive position is not. Revenue is cleanly ~96% PJM. Scarce nuclear baseload (Susquehanna) plus the AWS behind-the-meter PPA give real, hard-to-replicate optionality — but that is an asset/contract advantage, not market control. The oligopoly hard gate caps this at 5/10. Weight: 35%

AI / Data-Center Power Demand -- Strong Theme

Secular Tailwind -- Multi-Year Visibility

PJM is the epicenter of US data-center power demand. Summer peak load is forecast to grow ~3.6%/yr — +32 GW of peak load 2024-2030, ~30 GW of it from data centers — against a roughly flat-to-tight ~182 GW supply stack. That tightness is exactly why capacity-auction prices have set successive records and why scarce nuclear baseload is suddenly strategic. The theme is real and growing.

Price-Taker -- Not a Dominant Oligopolist

Oligopoly Gate: FAIL

Talen owns ~13 GW of a ~182 GW PJM market (~7% capacity share) and trails Dominion (~12%, 21.9 GW), Constellation (~11%, 20.3 GW), and Vistra (~8%, 13.9 GW). No single owner reaches ~12%; the top three control only ~30% combined, nowhere near 70%. Energy revenue clears at PJM LMPs and capacity revenue in the RPM auction — both set by the market/RTO, not by Talen. A price-taker in a fragmented field is, by the rubric, not an oligopolist.

Scarce Nuclear Baseload -- Optionality, Not a Moat

Susquehanna + AWS PPA -- Asset/Contract Advantage

The durable exception to commodity substitutability is the ~2.5 GW Susquehanna nuclear plant and its behind-the-meter PPA with AWS — carbon-free baseload co-located with a hyperscaler campus, hard to replicate because nuclear siting and interconnection are slow. But that is one PPA and an asset/contract advantage, not a market-share moat. Talen cannot dictate price or stop rivals (Constellation, Vistra, NRG, Dominion) from signing their own data-center PPAs.

Segment Revenue Mix (FY2025)

Segment	FY2025 Revenue	% of Revenue
PJM (Nuclear + Gas + Coal)	$2,477M	~96%
Other (ERCOT/WECC residual)	$161M	~6%
Corporate & elimination	— (net)	~(2)%
Total Operating Revenue	$2,581M	100%

By revenue type (FY2025): energy & other revenues $2,141M vs. capacity revenues $485M. Q1 2026 shows capacity revenue rising to $207M on energy & other of $1,034M — record PJM capacity-auction clears, but auction-cleared, not company-set, pricing.

Oligopoly Gate

Criterion	Result
TLN share of PJM capacity	~7% (~13 GW of ~182 GW)
Any segment above 30% share?	No
Top three control above 70%?	No (~30% combined)
Sets or takes prices?	Price-TAKER (LMP + RPM)
Key competitors	Dominion, Constellation, Vistra, NRG
Gate result	FAIL

5/10 — PJM data-center power demand is an A-grade secular theme with genuine multi-year visibility, and Talen's scarce nuclear baseload plus the AWS behind-the-meter PPA give real, hard-to-replicate optionality. But thematic exposure is not competitive position. Talen owns only ~7% of PJM capacity, trails Dominion, Constellation, and Vistra, and sets no prices in either the energy or capacity market. A big TAM and a roaring theme do not compensate for a fragmented, non-dominant position — the oligopoly hard gate applies and caps the score at 5/10.

Data sourced from Daloopa; PJM owner rankings and load forecasts per Utility Dive / Gabelli / PJM 2026 Load Report.