The AI gold rush has created a physical bottleneck: there simply isn't enough data center space and power. Anthropic's massive deal with Riot shows that Bitcoin miners — who already have the warehouses and electricity hooked up — are perfectly positioned

Verdict: wait — thesis has real legs but no confirmed entry

The idea that Bitcoin miners can monetize stranded power infrastructure for AI compute is genuinely compelling — Anthropic's reported $9B deal with Riot, per Cointelegraph, and surging TSMC sales cited by Investor's Business Daily confirm the demand side is real. CleanSpark's 102.2% revenue growth and 41.6% operating margin prove at least one company in the trio can fund itself, and the 24-month backtest's 26.6% return with a 54.7% win rate is respectable. But two of three stocks are structurally unprofitable — MARA's -20.9% operating margin and -39.2% revenue decline are jarring — and the basket's expected maximum drawdown of 97.3% under risk-parity is catastrophic. Right now every momentum entry condition is far from trigger across all three tickers: CLSK's RSI sits at 29.4 against a 50 threshold, its ADX at 11.1 against 25, and price is $1.75 below the Bollinger band. With a 48.0% drawdown in the full 60-month backtest and no robust parameter setup established, the setup simply is not live. **Conviction breakdown:** Thesis support is strong given real AI-demand contracts and CLSK's profitability. Trade readiness scores low because no entry condition cluster is close to triggering. Risk quality is weak given the near-100% modeled drawdown and a coin-flip 49.7% win rate. Backtest evidence is moderate — the recent window improved but the full-period drawdown is punishing. Fundamentals trend is mixed, with one profitable name carrying two deeply loss-making ones.

Trade now

**Do not buy today — every entry condition cluster is still far from trigger.** The strategy waits for a confirmed momentum reversal: price above the Bollinger (20) middle band, RSI (14) above 50 and crossing above 55, ADX (14) above 25, and RSI above the EMA (9). Right now all three tickers are in oversold or downtrending states, and none of the five entry conditions is met simultaneously on any symbol. **Live distance-to-trigger is large on every front.** CLSK at $11.74 needs to close above the Bollinger middle band at $13.50 — a $1.75 gap — while its RSI sits at 29.4 (needs above 50) and ADX at 11.1 (needs above 25). MARA is deeper in the hole: price $9.56 against a $10.92 Bollinger threshold, RSI 24.2, though its ADX of 83.8 already clears the trend filter. RIOT at $19.40 is closest on RSI (33.3, needs above 50) but still $1.43 below its Bollinger band and has the weakest trend reading at ADX 6.1. Only one condition — RSI above EMA (9) — is currently met across the trio, and that one alone is nowhere near enough. Once the setup does trigger, the risk frame is tight: a hard stop at 2.6% downside and a take-profit at 5.1% upside, yielding roughly 2:1 reward-to-risk before the momentum-based exit (RSI above 70 with price below the upper Bollinger band after five bars) or the 127.2% Fibonacci extension target comes into play. Position sizing is capped at 25% of portfolio per leg with 2.6% fixed-risk sizing against the second support level. "Wait" means exactly this: set alerts at the Bollinger middle band for each ticker and re-evaluate only when price reclaims that line with RSI pushing through 50. No robust parameter setup was established, so the published thresholds stand as-is. The backtest evidence supports the logic when conditions align — the 24-month CLSK window produced a 26.6% return with a 54.7% win rate across 75 trades — but that sample also endured a 35.6% drawdown, and the longer 60-month window's 48.0% drawdown underscores why the stop discipline matters. MARA's recent 24-month annualized return of -22.3% is a reminder that not every leg participates equally.

CleanSpark's profitability and Riot's AI pivot give the thesis real traction

The thesis hinges on Bitcoin miners monetizing their existing power and data-center infrastructure for AI compute. Per the Cointelegraph report on Anthropic's reported $9B compute deal with Riot, this is no longer hypothetical — a marquee AI customer is putting enormous capital behind the concept. Riot's 71.9% year-over-year revenue growth and relatively low debt-to-equity ratio of 0.21 suggest the company has the balance-sheet flexibility to pivot capacity toward higher-margin AI hosting without taking on prohibitive leverage. The idea argues that stranded mining infrastructure becomes some of the most valuable real estate in tech; that narrative has at least one major contract to…

Scores

• Conviction score breakdown: 40
• Thesis support: 72
• Trade readiness: 18
• Risk quality: 25
• Backtest evidence: 45
• Fundamentals trend: 40

Watch items

• CLSK — Price vs Bollinger (20) middle band
• CLSK — RSI (14)
• CLSK — ADX (14)
• MARA — Price vs Bollinger (20) middle band
• MARA — RSI (14)
• MARA — Annualized return (2-year)
• RIOT — ADX (14)
• RIOT — RSI (14)
• CLSK — Price above Bollinger (20)
• CLSK — RSI (14) above 50
• CLSK — RSI (14) crossed above 55
• CLSK — ADX (14) above 25
• CLSK — RSI (14) above EMA (9)