Financial Trends -- 8/10

Adj EBITDA tripled in four years -- the cleanest merchant power supercycle track record in the public set. Growth was front-loaded (+57% in 2022, +37% in 2023, +35% in 2024) and decelerated sharply to +5.4% in 2025 as hedge rolls and a milder Texas summer worked against the comp. 2026 guidance is $1.75-2.05B ongoing-ops EBITDA plus an incremental ~$500M run-rate from Cogentrix on close, indicating management sees the next leg coming from M&A and capacity revenue rather than spot price expansion.

GAAP revenue is noisy because it includes unrealized hedge MTM. The 2024 step-up reflects ~10 months of Energy Harbor consolidation; 2025 quarterly comps moderated. Q1 2026 re-accelerated to +43.4% YoY at $5,640M, suggesting realized power prices and merchant exposure are once again positive contributors heading into the heart of the data-center capex cycle.

East segment revenue +82% YoY in 2025 to $3,311M -- the Energy Harbor full-year contribution plus PJM capacity tightening. Retail (TXU + Ambit) grew +12% to $13.0B, confirming TXU continues to take share and capture price in ERCOT. VST does not formally report Vistra Vision vs Tradition revenue; the cleanest proxy is Retail (Vision) plus the wholesale Texas/East/West segments (mix of Vision nuclear and Tradition gas/coal).

Gas generation +54% from 2021 to 2025, capturing ERCOT load growth and PJM capacity demand. Nuclear is essentially flat at high capacity factors across Comanche Peak (ERCOT) plus Beaver Valley, Davis-Besse, and Perry (PJM, ex-Energy Harbor). Coal volumes are in managed decline (-11% from 2021). The mix shift from coal to gas + nuclear is exactly what the AI/data-center co-location thesis requires.

Two opposing forces -- shareholder-friendly buybacks against an aggressive debt-funded growth strategy. LT debt grew from $10.7B (2021) to $19.2B (Q1 2026, +79%) funding Energy Harbor (~$3B cash), buybacks, and Cogentrix-related capex. Interest expense hit $1,179M in 2025, +207% from 2021 and now ~20% of Adj EBITDA. But net leverage on Adj EBITDA actually fell from ~5.6x (2021) to ~3.0x (2025) -- VST is deleveraging on a multiple-of-EBITDA basis even as absolute debt rises, the strongest possible signal that growth capex is earning.

FCF compressed -47% in 2025 to $1.3B as capex stepped up to $2.75B -- the cycle bulge. Op cash flow remains structurally strong at $4-5B/yr; the FCF compression is entirely capex-driven (solar/storage development, growth gas at Vidor and Trinidad, nuclear uprates, Energy Harbor integration). Management's "ongoing operations adjusted FCF" -- the cleaner metric VST guides -- remains comfortably above $2B. Q1 2026 capex stepped up further to $883M, telegraphing continued growth investment.

VST's financial trends are the cleanest expression of the AI/data-center power-demand thesis in the IPP universe. Adj EBITDA tripled to $5.84B with a 4-year CAGR of ~32%; generation grew +20% from 2023-2025 to 208.2 TWh; share count is down ~28% via $6B of cumulative buybacks; GAAP operating income has been positive in every recent year; and the Energy Harbor acquisition cleanly integrated as East segment revenue +82% YoY in 2025. The dings are real but second-order: 2025 was a deceleration year (+5.4% Adj EBITDA after +35% in 2024); GAAP op income fell -53% YoY (rubric-mandated penalty though still positive at $1.9B); long-term debt nearly doubled to $19.2B and interest expense is now a meaningful EBITDA absorber; and capex is stepping up to compress implied FCF. Final score: 8/10 -- a premium to TLN's 6/10, justified by the longer track record, more durable cash generation, and shareholder-friendlier capital allocation.