Southern Company: 17 GW of data center demand funds a rate base bigger than the balance sheet can comfortably carry

Stevie AI on Southern Company (SO-USA | southerncomp)

8/31/2026

Summary

Southern Company is a vertically integrated, state-regulated electric utility serving over 9 million customers across Alabama, Georgia, and Mississippi through Alabama Power, Georgia Power, and Mississippi Power, supplemented by Southern Company Gas and Southern Power. The structural insight for 2026-2029 is that Southern has effectively pre-sold its next leg of growth: 17+ GW of contracted large-load and data center demand (including the OpenAI deal, backed by $20.8 billion in aggregate collateral at full ramp) locks in rate-base expansion with cost-to-serve protections, while retail base rates in Georgia and Alabama stay flat through 2028-2029. That combination is unusually clean regulatory economics — the utility earns on the incremental capex the load requires without reopening politically sensitive general rate cases. But the financing math is the offsetting reality: net debt climbs from roughly $79.8 billion in FY2026 to $102.4 billion by FY2029 against consistently negative free cash flow of $3-3.5 billion annually, funding a $13-15 billion/year capex program. FY2025 actual results show net income of $4.3 billion and EPS of $3.92, down from $4.4 billion and $3.99 in FY2024 — a modest step back that reflects share count growth and cost pressures ahead of the load ramp, not underlying earnings power deterioration. Management's guidance of $4.50-$4.60 adjusted EPS for 2026, with confidence in landing at or above the top of that range, implies a reacceleration versus the GAAP EPS path shown in the forecast ($3.79 in FY2026, rising to $4.41 by FY2029), a mix that reflects both adjusted vs. GAAP presentation differences and the phasing of rate base earnings against elevated interest expense. We apply an 18x forward P/E to our EPS estimates, consistent with Southern's historical trading range as a premium-regulated utility with above-average rate base growth visibility, but not at a premium to best-in-class peers given the financing overhang. Applying 18x to FY2026-FY2029 EPS of $3.79, $4.07, $4.30, and $4.41 yields price targets of $68.22, $73.26, $77.40, and $79.38 — all below the current price of $88.17. At today's price, the stock is discounting a smoother execution and financing path than the numbers currently support, leaving limited near-term upside and skewing the risk/reward toward patience rather than accumulation.

Thesis

1. **Regulated Monopoly Economics with a Structural Growth Kicker**: Southern's core advantage is that it operates exclusive service territories in the Southeast with essentially no retail competition, and it has now layered an unusually large contracted demand pipeline (17+ GW) onto that regulatory base. Unlike prior utility growth cycles driven by speculative capacity additions, this buildout is pre-contracted with minimum-bill provisions covering 100% of incremental cost to serve, materially de-risking the revenue side of the capex program relative to a typical utility expansion. 2. **Rate Stability as a Political and Financial Asset**: Holding Georgia and Alabama retail base rates flat through 2028-2029 — even while absorbing 17+ GW of new large load — is a deliberate strategy to preserve regulatory goodwill. Projected annual customer savings of $950 million beginning in 2029 (roughly $180/year for a typical Georgia Power residential customer) gives management a credible public narrative that large-load customers are subsidizing, not burdening, the retail base. This reduces the risk of political backlash or rate case litigation that has hit other utilities pursuing large capex programs. 3. **EPS Growth Skewing to Top of Algorithm, But Off a Downgraded Base**: FY2025 EPS of $3.92 was down from $3.99 in FY2024, and FY2026 GAAP EPS in our forecast is $3.79 before recovering to $4.07 in FY2027 and $4.41 by FY2029. Management's adjusted guidance of $4.50-$4.60 for 2026 sits well above this GAAP figure, and the credibility of "top-of-range" delivery will be tested by Q3 2026 results (guided to $1.50/share). The reacceleration story is real but back-loaded — most of the earnings benefit from the load buildout doesn't show up meaningfully until FY2028-FY2029. 4. **Financing the Capex Program Without Buybacks**: With FCF negative $3.1-3.5 billion annually through FY2029 and net debt rising from $79.8 billion to $102.4 billion, Southern is explicitly prioritizing debt and "modest equity issuance" over shareholder returns via buybacks. This is a reasonable trade-off given contracted revenue visibility, but it means EPS growth must come almost entirely from rate base expansion outpacing dilution and rising interest expense — a narrower margin for error than a utility funding growth from internally generated cash. 5. **RFP Pipeline Extends the Growth Runway, But Adds Uncertainty**: Georgia Power is reportedly 1-2 GW away from exhausting approved generation capacity following the OpenAI contract, and new RFP-selected resources (selection by end of 2026, certification by late 2027) represent "substantial incremental investment" beyond the current base capital plan. This is upside optionality to the growth algorithm, but it also means the current $13-15 billion/year capex figure likely understates the eventual run-rate, with financing implications not yet reflected in consensus. 6. **Valuation Gap Signals the Market Is Ahead of the Financing Reality**: At $88.17, the stock trades well above the 18x multiple applied to our FY2026-FY2029 EPS path ($68-$79 range), suggesting the market is pricing in either faster-than-modeled adjusted EPS growth, a lower cost of capital than the rising net debt trajectory implies, or a re-rating toward growth-utility multiples that we view as premature until the RFP-driven capex layer and its financing are clarified.

Risks

1. **Execution Risk on an Unprecedented Capital Buildout**: Deploying 10+ GW of new generation plus hundreds of miles of transmission over 5-7 years, on top of ongoing base capex, creates meaningful risk of cost overruns or schedule slippage, which could require incremental equity issuance beyond current 'modest' assumptions and pressure both dilution and leverage targets. 2. **Rising Leverage Without Offsetting Free Cash Flow**: Net debt is forecast to grow by roughly $22.6 billion from FY2026 to FY2029 while FCF remains negative every year in the forecast window. Should interest rates stay elevated or credit spreads widen, rising interest expense could erode more of the rate-base earnings benefit than currently modeled, directly pressuring the EPS growth algorithm. 3. **Large-Load Demand Concentration and Counterparty Risk**: The growth thesis leans heavily on a relatively small number of large contracted customers (e.g., OpenAI). While minimum-bill provisions and collateral requirements (A- or better credit standards) mitigate this, any restructuring, delay, or reduction in hyperscaler/data center capex plans industry-wide could slow the load ramp that underpins the entire rate base growth narrative. 4. **Regulatory Reversal Risk on Rate Flatness**: Keeping retail rates flat through 2028-2029 is a political strategy that depends on large-load customers actually delivering the promised cost offsets. If large-load revenue underperforms or costs to serve run higher than modeled, regulators could face pressure to revisit the arrangement, or Southern could face margin compression absorbing the gap. 5. **Generation Capacity Constraints Requiring Unbudgeted Investment**: Georgia Power being only 1-2 GW from exhausting approved capacity means the RFP process (selection end-2026, certification late-2027) is not optional — new resources are required to serve contracted load. Delays in certification or unfavorable RFP economics could either delay revenue recognition on committed load or require additional capital beyond the current plan with uncertain returns. 6. **Valuation Disconnect Versus Modeled EPS and Multiple**: Our 18x multiple applied to modeled EPS implies targets of $68-$79 across FY2026-FY2029, all below the current $88.17 price. If the market continues to award a growth-utility premium multiple (above 20x) independent of the financing overhang, the stock could remain range-bound or de-rate sharply on any capex, rate case, or large-load disappointment.

📈 Price Targets