VST — Vistra Corp.

The sell-side is firmly bullish: 20 Buy versus 2 Hold ratings, no Sells, with consensus targets of $217-228 implying roughly 45% upside from the $148.38 close on 2026-09-13. The narrative is the AI/data-center power build-out - the August 2026 Helix Digital Infrastructure venture with KKR, the Kuwait Investment Authority and NVIDIA, plus prior nuclear agreements, and favorable PJM capacity outcomes. The tape is more cautious than the commentary, though: as of 2026-09-13 the stock was below its 200-day SMA, down 6.65% over six months, and the ~$50B valuation now leans on executing that narrative rather than on delivered GAAP earnings (TTM operating margin of just 2.3%).

Revenue is noisy and recently declining year over year: Q2 2026 $4.02B vs. Q2 2025 $4.25B (-5.5%), Q1 2026 $4.65B vs. Q1 2025 $5.17B (-10.0%), and Q4 2025 collapsing to $2.34B from $7.36B a year earlier - swings that reflect hedging/pass-through and portfolio changes rather than demand, which is why the deterministic model flags two >30% structural revenue breaks. GAAP operating income is even more erratic: Q4 2024 was $3.81B and Q3 2025 $1.04B, then Q4 2025 printed an operating loss of -$1.748B (gross profit -$1.301B, consistent with a large unrealized hedge/mark-to-market charge like the $472M unrealized hedging loss booked in Q2 2026). It has since recovered to $527M (Q1 2026) and $553M (Q2 2026). The structural read: reported GAAP earnings are compressing and volatile while ongoing-operations Adjusted EBITDA is accelerating (+30%+ YoY to $1.767B in Q2 2026) with full-year guidance reaffirmed - the underlying cash business is inflecting up, but reported earnings quality is poor and hard to model. Free cash flow is similarly lumpy: $133M in Q2 2026, $316M in Q1 2026, -$82M in Q4 2025 and $1.009B in Q3 2025, against heavy capex (~$890M in Q2 2026).

Mixed-to-fair. As of 2026-09-13 the stock trades at 24.8x TTM EPS of $6.53 but only ~17.2x next-FY consensus EPS of $8.61 - a big step down that assumes ~32% EPS growth. Versus its own history that forward number looks cheap (3-year average P/E of 34.3x), but it is expensive on book (9.2x P/B) and sales (3.13x P/S vs. a 5-year average of 1.59x), and EV/EBITDA of ~10.8x is a full multiple. Net: roughly fair - you pay a moderate multiple on forward numbers for a levered, cyclical power producer whose premium depends on data-center demand actually showing up.

The stock is priced for the earnings inflection, not the trailing print. The leap from $6.53 TTM EPS to $8.61 next-FY consensus - and analyst targets of $217-228, roughly 25-26x forward EPS - assumes data-center PPAs, favorable PJM capacity pricing and the Cogentrix deal all land. If they do, today's ~17x forward P/E is genuinely cheap and the multiple can hold; if guidance slips or the GAAP-vs-adjusted gap persists (Q4 2025's -$1.75B operating loss is the cautionary case), a 9x book multiple leaves real downside. Watch the Q3 2026 print for the data-center contract pipeline and any revision to the $6.8-7.6B EBITDA range.

As of 2026-09-13, RSI(14) stood at 54.5 - neutral, neither overbought nor oversold. Price sat +3.95% above the 20-day SMA (~$142.7) and essentially on the 50-day (~$148.9, -0.32%), but -5.66% below the 200-day (~$157.3), with a 6-month return of -6.65% and a 52-week range of $132.66-$219.82. That is a range-bound consolidation below the long-term trend - not a broken chart, but not a confirmed uptrend either. Timing implication: no urgency to chase; a reclaim of the ~$157 200-day would confirm the trend, while a loss of the 50-day risks a retest of the $132-135 area.

• AI/hyperscale data-center electricity demand is structurally tightening U.S. power markets - Vistra's nuclear and gas fleet and its preferred-supplier role in the Helix venture (with KKR, KIA and NVIDIA) are direct beneficiaries.
• PJM capacity auction outcomes and wholesale/locational power pricing set the ceiling on unhedged cash generation; strong recent auctions have lifted cash flow but are politically contentious and subject to FERC market-design reform.
• Interest rates and the cost of capital: the September 10, 2026 pricing of $1.5B of junior subordinated notes at 7.00%/7.25% shows refinancing is not cheap, and Vistra is funding growth and preferred redemptions with long-duration debt.
• Regulation/FERC: market rules, capacity-market reform and merger review (Cogentrix, Helix) are live swing factors on both growth and returns.
• Commodity and weather risk: natural gas prices and weather drive unhedged power economics; near-term hedging (~100%) caps upside as well as downside.

• Settlement of the $1.5B junior subordinated notes on September 24, 2026, funding preferred-stock redemptions at the October and December 2026 five-year reset windows - a balance-sheet simplification signal.
• Closing of the FERC-approved Cogentrix Energy acquisition in H2 2026, which adds generation scale to the data-center thesis.
• Q3 2026 earnings in early November 2026: updated data-center contract pipeline, hedging position and precision on the $6.8-7.6B EBITDA / $3.925-4.725B FCFbG guidance.
• Additional data-center PPAs (following Meta and the Helix formation) that would convert the AI narrative into contracted, financeable cash flow.

• Leverage and refinancing risk: $17.7B of long-term debt against only $5.48B of equity (D/E 3.63x), a 0.97 current ratio, and a 7%+ marginal cost of new subordinated debt.
• Earnings-quality opacity: Q4 2025's -$1.748B GAAP operating loss and TTM operating margin of 2.3% sit in stark contrast to adjusted figures - a persistent GAAP/adjusted gap makes results hard to trust and can trigger sharp re-ratings.
• Power price and hedge risk: near-fully hedged near-term exposure limits upside if wholesale prices spike, while mark-to-market swings (e.g. the $472M unrealized hedging loss in Q2 2026) can whipsaw reported earnings.
• Narrative risk: the valuation premium rests on AI/data-center demand; an AI capex slowdown or competitive IPP response would compress the multiple quickly.
• Technical/momentum risk: the stock is below its 200-day SMA and has a negative six-month return, leaving it vulnerable to continued de-rating if Q3 disappoints.

Vistra is a genuinely improving cash-flow story - reaffirmed $6.8-7.6B 2026 EBITDA, ~$3.9-4.7B adjusted FCFbG (a high-single-digit yield on today's ~$50B cap), and prime positioning for the AI/data-center power build-out - trading at a reasonable ~17x forward EPS and a discount to its own 3-year average P/E. The catch is that it is a levered (D/E 3.6x, $17.7B LT debt) business whose GAAP earnings are opaque (Q4 2025 operating loss of -$1.75B) and whose stock sits below its 200-day after a 6.65% six-month decline. Reasonable to own on the forward numbers, but I would size modestly and want either a reclaim of ~$157 or a pullback toward the low $130s as a better entry.

Blended multiple approach: ~20.7x next-FY consensus EPS of $8.61 yields ~$178, corroborated by ~10.7x 2026 guided Adjusted EBITDA of ~$7.2B (EV ~$77B, less ~$17.3B net debt, divided by ~337M shares = ~$177) and cross-checked against a ~7.5% fair adjusted-FCF yield on the $4.3B FCFbG midpoint (~$171). I used a mid-teens-to-low-20s multiple - below the 3-year average P/E of 34.3x, which is distorted by depressed trailing GAAP earnings - to balance ~30% adjusted EBITDA growth against above-average leverage and GAAP/adjusted earnings opacity.