$MARA — the framework I use for Marathon Digital. Reference price: $16.80 | FY26 rev: $1.2B | NI: $300M | Shares: 215M 5-year bull / base / bear model: Bear Base Bull 2031 price $5–14 $12–23 $50–75 Annual return ~-15–20% ~-5–0% ~24–35% Equity value $1–3B $2.5–5B $11–16B FCF multiple 8× 12× 18× Probability 35% 45% 20% Bear case for MARA (35% probability): BTC drops below $50k, halving margin compression hits hard, AI pivot fails. Revenue $600-800M, FCF goes negative. Stock trades near cash. The risk here isn't small. Marathon Digital has real structural problems that won't resolve quickly. Base case (45%): BTC consolidates $60-90k. Hash rate grows, efficiency improves. Revenue $1.4-1.8B, FCF $200-400M. Stock trades as direct BTC beta. This is the "nothing special, nothing terrible" scenario. MARA survives and muddles through. Bull case (20%): BTC hits new highs $120-150k, MARA holds hash rate share, AI compute deals materialize. Revenue $2.5-3B+, FCF $600-900M+. Direct lever on BTC price. This is where the real upside lives. But it requires everything to go right. The critical assumption The entire thesis rests on: BTC holds above $55k AND the AI compute pivot lands real contracts. Below $50k BTC, the math breaks for most miners. Numbers to watch over the next 4-6 quarters: • Cost per BTC mined (hashprice) • Hash rate growth vs competitors • AI/HPC compute contract announcements • Debt levels and equity raises My rough probability weighting: Bear 35% / Base 45% / Bull 20% Rough probability-weighted value: ~$19.00 The interesting exercise: what does Marathon Digital look like at 30-40% below current prices? Because at that level, the risk/reward changes dramatically — even if the bear case plays out. MARA — high conviction name if you believe the thesis. Value trap if you don't.