Vistra Stock Sits 37% Below Its High While Power Demand Keeps Climbing. Should You Buy It?

Vistra reported Q2 adjusted EBITDA from ongoing operations up 31% YoY to $1.77B, with 2026 adjusted EBITDA guidance reaffirmed at $6.8B–$7.6B (management targeting at/above the midpoint). The company expects $3.9B–$4.7B of adjusted free cash flow this year and said ~100% of 2026 generation volumes are hedged, while adding long-dated nuclear power contracts with Amazon Web Services (1,200 MW, deliveries late-2027) and Meta (2,609 MW, deliveries starting late this year). Despite the shares being down ~37% from the $219.82 52-week high to about $139, the article frames the drawdown as valuation/AI-power crowding rather than deteriorating fundamentals, noting the forward P/E is ~13.

Analysis

The market is still pricing VST as a crowded AI-power beta trade, but the business is migrating toward long-duration contracted cash flow with residual merchant upside. That matters because the near-term catalyst is not earnings; most of the 2026 volume is already economically spoken for, so the stock can stay range-bound even if reported results remain strong. The bigger rerating lever is whether the market starts capitalizing the backlog like infrastructure rather than commodity exposure.

Second-order, AMZN and META are not just customers; they are signaling that hyperscalers are willing to lock baseload supply years ahead, which should improve bargaining power for other nuclear owners and eventually tighten the multiple spread between contracted names and merchant-heavy generators. CEG is the cleanest public comparator, but the broader loser group is any power producer that can’t offer low-carbon firm capacity or interconnection certainty. On the bond side, durable PPAs should help credit spreads over time, especially if management keeps buying back stock instead of levering up for growth.

The contrarian risk is that the market is probably underestimating timing risk. The demand story is real, but the equity only captures it if data-center capex and grid buildout stay on schedule; a six-month pause in hyperscaler deployment would hit sentiment long before it hits cash flow. Also, hedge-accounting noise can keep headline earnings volatile and mask the underlying FCF story, which means the stock can stay cheap longer than bulls expect.

What would falsify the thesis is not a weak quarter; it would be a downgrade to 2027 EBITDA or evidence that new long-term contracts stop clearing at attractive prices. If that happens, the valuation argument breaks and this becomes just another cyclical power name.

Key Decisions for Investors

• Long VST on weakness, not strength; use pullbacks into the $130s with a 3-6 month horizon. Base case is a rerate toward a higher multiple as the market focuses on contracted cash flow rather than near-term hedge noise; thesis breaks on a sustained move below ~$120 with no guide raise.
• Pair trade: long VST / short CEG, sized market-neutral. VST screens cheaper on forward cash flow and appears more visibly contracted; if the AI-power trade cools, the higher-multiple nuclear peer should compress faster.
• Watchlist alert: if VST lifts 2027 EBITDA guidance or signs another multi-year hyperscaler PPA, add to longs. That would be the first real catalyst for a structural multiple expansion over the next 1-3 months.
• Avoid chasing AMZN/META for the power narrative alone; the benefit is strategic, not earnings-driving. Treat any upside there as a secondary consequence of data-center optionality, not a direct trade.
• If implied volatility remains elevated, consider VST call spreads instead of outright stock for a 6-12 month expression; the thesis depends more on rerating than near-term earnings acceleration.