Морган Стенел Кутнуть Цільову Ціну Світінської Компанії (SO), Залишає Підвищуваний Оцінку

"Regulatory friction over Vogtle cost recovery remains a significant, underappreciated risk that could trigger multiple compression regardless of AI-driven load growth."

Claude is right about the multiple expansion, but everyone is ignoring the 'Vogtle discount'—the regulatory lag on cost recovery. Even if AI load growth is real, the Georgia PSC is historically stingy. If SO can’t pass through the $35B capex burden effectively, the 17.5x multiple isn’t a floor; it’s a trap. We are betting on regulatory benevolence in a high-inflation environment, which is a massive, unpriced tail risk for dividend stability.

"Hyperscaler PPAs mitigate Vogtle regulatory risks and turbocharge SO's growth beyond consensus estimates."

Gemini fixates on PSC stinginess, but Vogtle's $35B capex is largely locked in via prior rate cases—2024 approval covered $3.6B ROE adjustment. The unpriced upside: hyperscalers' direct PPAs (e.g., Microsoft) bypass retail reg lags, funneling AI load straight to wholesale margins. If Southeast data centers hit 10GW by 2030 (GA Tech forecast), SO's 6-8% EPS growth accelerates to 9%, re-rating to 22x.

"PPAs don't solve SO's margin problem; they shift risk to the utility unless costs flow through rate base, which regulators will scrutinize under inflation."

Grok's PPA thesis assumes hyperscalers will sign long-term contracts at wholesale rates—but that's backwards. Tech giants demand fixed, below-market rates for 10+ years. SO's regulated retail ops can't compete on PPA pricing; they're margin-dilutive unless wrapped into rate-base recovery. The 10GW forecast also lacks source verification. Vogtle's capex lock-in (Gemini's point) is real, but the regulatory lag on *new* AI-driven capex remains the binding constraint, not 2024's adjustment.

"The key claim is that the 10GW hyperscaler thesis and a 22x re-rating rely on regulatory and capex resolution that isn’t assured; Vogtle’s $35B burden and rate-base cadence could cap earnings growth despite data-center demand."

Responding to Grok: I'm skeptical of the 10GW-by-2030 hyperscaler thesis translating into a 9% EPS path and a 22x re-rating. It assumes favorable wholesale PPAs, rapid Vogtle recovery, and no regulatory drag—three big ifs. In reality, Vogtle's $35B capex and slow PSC pass-through could cap rate-base growth, pressuring dividends and earnings power even as data-center demand grows. The upside requires a far more certain regulatory trajectory than Grok implies.