When: Daily close below 91.80 (200 SMA)

Then: Trim position defensively

HOLD

• Forward P/E of 18.82 on $4.92 EPS consensus estimate represents full valuation with limited margin of safety; PEG of 2.36 is not cheap for a regulated utility
• Technical deterioration: MACD histogram expanded from negative 0.13 to negative 0.41, RSI fell from 55 to 40, ATR expanded 41%, price sitting 1.72 below 50 SMA with three of eight indicators bearish
• Interest expense grew 60% over three years from $2.02 billion to $3.24 billion; higher-for-longer Fed rates (86% probability on Polymarket) continue to pressure this
• Negative free cash flow in three of last four years; Q2 2026 required $2.06 billion equity issuance, causing shareholder dilution
• 73% probability at least one state enacts data center moratorium this year (rising 2.3 percentage points weekly), creating contagion risk for Georgia PSC approvals
• Fundamentals improving: 17 gigawatt pipeline, OpenAI contract, 55% YoY data center demand growth, but these are already priced in at current levels
• Dividend yield of 3.27% with 31% payout ratio on operating cash flow provides income support while awaiting better entry
• Recommended stop at 91.80 (200 SMA), but ATR of 2.05 suggests widening to 91.50 or committing strict execution discipline; reduce add-on allocation from 25% to 15-20% on breakout above 94.41

→ vs aggressive: The Aggressive analyst's 'generational setup' framing identifies a move that has already happened; stock went from low 80s to high 90s in twelve months, rationalizing entry at the top

→ vs aggressive: Forward EPS of $4.92 is analyst consensus, not delivered result; assumes rate base returns on schedule, no Plant Bowen cost overruns, no ROE haircuts from Georgia PSC, and no further interest expense growth

→ vs aggressive: Bear case requires more than 'one thing' - it's a portfolio of execution risks: OpenAI Savannah execution, hyperscaler pace, PSC rate cases, $12.7B capex without overruns, benign macro; not asymmetric favoring bulls

→ vs aggressive: The 73% data center moratorium probability is systemic contagion risk, not just regional political risk; once one large state moves, others accelerate under political pressure

→ vs aggressive: Revenue miss despite EPS beat signals quality-of-earnings divergence; top line weaker than bottom line suggests, worth respecting not dismissing

→ vs aggressive: FCF situation worse than acknowledged; negative FCF in three of four years requires continued debt issuance or equity dilution ($2.06B in Q2 2026), diluting existing shareholders

→ vs aggressive: Claim that 'no Fed cuts already priced in' is self-defeating - if priced in it's neutral, not a catalyst; can't claim both no impact and upside potential

→ vs aggressive: June 1 RSI bottom of 32.74 is being used backwards - that occurred AFTER correction, not as a leading indicator of current momentum deterioration

→ vs neutral: Patience-over-prediction argument is closer to conservative playbook than Aggressive analyst admits; trader's HOLD with defined parameters is appropriately conservative

HOLD

• HOLD decision stands - closer to trader's original position than either extreme
• Refinement 1: Reduce add-on trigger from 25% to 20% to account for valuation premium
• Refinement 2: Leave stop at 91.8 as a closing-price stop, not intraday - preserves structural line without ATR noise
• Refinement 3: Consider covered call writing against existing position to harvest elevated implied volatility
• Breakout confirmation requires three simultaneous conditions: daily close above 94.41, positive MACD histogram, and RSI above 50
• Three-indicator confirmation on stock with beta 0.327 is a reasonably tight filter
• ATR has expanded 41% - historically corresponds to elevated implied volatility for covered call income
• Covered call at strike near 95 or 96 would generate income while capping upside
• No retail engagement data - institutionally driven sentiment means moves come from news flow or macro catalysts, not retail momentum
• 91.8 is the structural line - closing below it breaks the long-term thesis regardless of whether it was noise
• Dividend at 3.27% yield with 31% operating cash flow payout ratio is not at risk
• Downside is fast and discrete; upside is slow and conditional - asymmetry favors defensive position sizing

→ vs aggressive: Generational setup language causes investors to anchor to thesis and ignore new information - stock already moved from low 80s to high 90s in 12 months

→ vs aggressive: Forward EPS of $4.92 is a consensus estimate, not delivered result - assumes portfolio of execution successes

→ vs aggressive: PEG of 2.36 is a premium metric, not a value metric - cannot hand-wave away trailing and forward multiples both being above historical averages

→ vs aggressive: Macro argument is confused - cannot simultaneously claim no cuts is priced in AND dovish surprise would cause 3-6% rally; Polymarket has no-cut probability rising 2.5 percentage points week over week, macro tape is leaning hawkish

→ vs aggressive: Betting on low probability dovish event while accepting high probability outcome as neutral is asymmetric risk against, not in favor

→ vs conservative: Geographic distribution of moratorium risk is critical - SO operates in Georgia, Alabama, Mississippi, not Oregon/Virginia; Georgia governor courting hyperscaler investment, Mississippi/Alabama need industrial capex - moratorium narrative driven by Pacific Northwest water stress, not Sunbelt states

→ vs conservative: Stop refinement suggestions of 91.5 or 91.8 with strict execution are not equivalent - 91.8 with closing stop does not trigger on intraday wicks, but 91.8 with intraday stop does get shaken out by ATR expansion as shown on August 7 when low tagged 91.75 and closed at 92.69

→ vs conservative: Trim into rallies makes no sense when dividend is fully covered at 31% operating cash flow payout ratio - income stream is not at risk

→ vs conservative: Correctly flagged FCF situation and $2 billion equity dilution in Q2, but conservative framing of these risks is appropriate

→ vs conservative: Add-on trigger of 15-20% underestimates the reliability of three-indicator confirmation combined with low beta 0.327 - 20% is the appropriate middle ground