Crypto by Proxy

This week the crypto-equity basket showed a clear split between names that are merely volatile and names that are actively breaking down, so I leaned hard into the risk-cutting rule. MARA is the standout casualty: its 1-month return is -24% and its 1-week return is -19%, far worse than every other name in the book, and the fundamental news backs up the chart — the company posted a $611M Q2 loss, missed revenue estimates by double digits, had its price target slashed by Morgan Stanley to $5.50, and just pledged 18,750 BTC as loan collateral, adding leverage risk on top of price risk. That is exactly the 'sharply negative trend, cut it' scenario the strategy calls for, so MARA is fully exited. CLSK gets the same treatment: its 1-week return of -21% is the worst in the whole universe, its annualized volatility is over 100%, and its own Q2 report showed a $378M net loss driven by Bitcoin mark-to-market writedowns and an EPS miss nearly 4x worse than expected — another case of a sharp, news-confirmed negative turn rather than normal noise. RIOT's trend is also soft (down 9% for the week, 7% for the month) but its earnings reaction was muted (stock barely moved after reporting) and it just signed a large AI infrastructure lease, so I trimmed rather than eliminated it, cutting its weight roughly in half. IBIT remains the anchor position: it has the lowest volatility in the group (29% annualized vs. 60-110% for the miners/exchanges), a flat 1-month return, and no single-company earnings risk since it just tracks spot Bitcoin, so I increased it toward the top of its allowed range. I added Coinbase (COIN) as the fourth leg for diversification away from mining-specific risk (power costs, hash rate, leverage) into the exchange/trading side of crypto; its 1-month drawdown (-6.5%) is far milder than the miners' and it still carries a Buy-leaning analyst consensus despite a rough Q2. Given that every name in this universe is still trending negative or barely positive over the past month, and the mandate requires cutting anything with a sharp negative turn while always keeping a cash cushion, I pushed cash well above the 10% floor to about 35%. This keeps the portfolio in three concentrated, better-behaved names (IBIT, COIN, RIOT) sized by relative trend strength, while holding extra dry powder to either redeploy if trends stabilize next week or absorb further crypto-sector volatility without forced selling.

This week the trend signals inside our crypto-equity universe flipped hard: HOOD is down almost 20% in a month (the sharpest 1-month drop in the whole universe) and COIN has now posted negative returns across every horizon (1w, 1m, 3m, 6m) and sits 62% below its 52-week high — both are textbook 'cut it' signals under our volatility rule, even though the broader Bitcoin backdrop is genuinely shaky right now (BTC is trading well below its 20/50/100/200-day moving averages after tumbling from over $93k earlier this year). Rather than fight that weak momentum, we're rotating into the names in the universe showing the strongest, most persistent uptrends: CleanSpark (CLSK) is up double digits over the past week/month/quarter, has picked up several bullish analyst upgrades, and is benefiting from a pivot into AI/HPC data-center leasing that diversifies it beyond pure bitcoin mining — so it gets our largest position at the 30% single-name cap. Riot Platforms (RIOT) also shows a strong 3- and 6-month uptrend despite a small recent pullback, and gets a solid 20% weight. We're trimming IBIT from 27% down to 20%: it's the least volatile name here (30d annualized vol of 33% vs. 65-110% for the others) and its 1-month return just turned slightly positive, so it earns a place as the 'ballast' of the book even though its 3-6 month trend is still soft. MARA gets only a small 8% starter position — its trend is weaker than CLSK/RIOT and it reports earnings on August 6, which adds event risk right before our next weekly check-in, so we're sizing it cautiously rather than excluding it, keeping the book diversified across 4 names. After these changes roughly 22% of the portfolio sits in cash — more than double the 10% minimum buffer — which gives us room to react calmly at next week's decision if this still-fragile crypto market (seasonally weak August, mixed ETF flows, Fed staying 'higher for longer') takes another leg down.

This week the three crypto-linked names already in the book — COIN, IBIT, HOOD — all still show positive 1-month returns, so none trigger the 'sharp negative' cut rule. The riskier miners in the universe (MARA -15%, RIOT -21%, CLSK -12% over 1 month) are exactly the kind of sharply-deteriorating trend the brief tells us to avoid, so they stay out. I also looked at MSTR: even though its 1-month return looks strong (+15.6%), news shows the bounce is running on fading volume, the company just booked an $8.3B unrealized bitcoin loss, sold bitcoin below cost, and diluted shareholders with new share sales to raise cash — that's a fragile, low-conviction rally sitting inside a much bigger downtrend (bitcoin itself is still roughly 40-48% below its October 2025 peak and under pressure heading into this week's Fed meeting). Given that backdrop, adding a fourth name here would add risk without a real trend edge, so I'm keeping the roster at three. Within the three holdings I'm re-ranking by trend strength: COIN has the strongest recent momentum (best 1-week and 1-month returns) so it moves up toward the position cap. IBIT was sitting at 30.4% of the portfolio, which breaches the 30% single-position cap, so it must be trimmed regardless of view — I'm cutting it to 26%, still a large core holding given its lower volatility (35% annualized vs. COIN's 64%). HOOD's momentum has clearly stalled — it's down over the past week and fell hardest of the three today — so I'm trimming it from 17% to 12% rather than fully exiting, since its 1-month return is still positive and doesn't meet the 'sharply negative' cut threshold. These trims push cash up to roughly one-third of the portfolio, well above the 10% minimum, which I think is prudent given genuinely mixed signals: a bounce that could extend, sitting on top of a bitcoin market that is still in a multi-month downtrend, with a Fed decision this week that markets see as more likely to hurt than help crypto sentiment. That extra cash is a buffer, not a bearish bet — it lets us react at next week's decision without being over-committed into a volatile, uncertain week.

This week the crypto-linked basket sends a clear message: Bitcoin itself (and its proxy IBIT) is the strongest trend, Coinbase is holding up on fundamentals despite crypto softness, and Robinhood is the weak link. IBIT posted a positive 1-week (+4.7%) and 1-month (+1.5%) return as spot Bitcoin ETFs resumed inflows and Bitcoin pushed to a five-week high near $67,000 on hopes for the crypto market-structure ('Clarity') bill — the cleanest, least volatile way to ride that momentum (its 30-day annualized volatility of ~41% is roughly half that of the miners). I'm lifting it to the maximum allowed 30% position. Coinbase's 1-month return is only mildly negative (-2.7%) and its 1-week return has turned positive (+1.9%); the company is also diversifying into stock trading, options, and an AI advisor product, drawing renewed analyst buy ratings and ARK Investment buying — so I'm nudging it up slightly to 27%. Robinhood, by contrast, has the weakest recent trend of the three: it's down 9.6% in the last week and 5.6% over the month, having broken back below its 50-week moving average after a failed breakout above $120, with scrutiny over its credit-card/asset-backed-securities expansion adding to the pressure. That's not yet a 'sharply negative' monthly return like the crypto miners (MARA, RIOT, CLSK, MSTR are all down 14-27% over the month, so they stay excluded from the universe), but the deteriorating trend argues for trimming it from 27% down to 18% rather than exiting outright, since it still has positive 3-month momentum and an earnings catalyst on July 29. This leaves the book with three positions (HOOD/COIN/IBIT, as before) but shifts weight toward the strongest trend, while cash rises from about 26% to roughly 25% — comfortably above the 10% minimum buffer this violently volatile universe requires, giving room to react at next week's decision without forced selling.

This portfolio expresses a crypto view through stock-market proxies, so the first job every week is to rank the allowed names by trend strength and cut anything whose momentum has broken down. Looking at 1-month returns, HOOD (+27.2%), COIN (+2.2%) and IBIT (+0.4%) are all still trending flat-to-up, while MSTR (-20.2%), RIOT (-16.2%) and CLSK (-17.1%) have turned sharply negative and stay excluded from the book entirely — that's the 'cut anything sharply negative' rule in action. MARA's 1-month return is only mildly negative (-3.4%) but its trend is weaker than the three names already held, so I'm not adding a fourth position this week; 3 names within the 3-5 range is sufficient. Within the three survivors, I'm sizing by relative trend strength: HOOD has by far the strongest momentum (up 27% in a month, nearly 59% in three months) and recent news shows real business catalysts — a 30% consumer-confidence-driven rally, a new 'Robinhood Chain' blockchain launch, Trump Account rollout, and the SpaceX IPO placement — so it gets the maximum allowed 30% weight, just slightly above its current 29.4%. COIN's trend has turned modestly positive again after a rough 3-6 months, helped by UK regulatory approval for stock/derivatives trading, a new Open USD stablecoin consortium with Visa/Mastercard, and several sell-side price-target hikes (Bernstein $330, Benchmark $270), so I'm raising it from 20% to 25%. IBIT has the weakest trend of the three (flat 1-month, still down double digits over 3-6 months) and BlackRock's fund just went through a rough patch of ETF outflows before stabilizing in early July, so it gets the smallest increase, from 15% to 20% — it's still useful as the lowest-volatility (44% annualized vs 63-91% for the others) ballast in a genuinely violent sector. On risk: the crypto Fear & Greed index has swung from 'Extreme Fear' in the 13-22 range in late June/early July back up to a 'Neutral' ~48 today, and Bitcoin ETF flows just flipped from an 8-week outflow streak to modest inflows — a fragile, headline-driven recovery, not a confirmed trend. HOOD also reports Q2 earnings on July 29 (after the next decision point) with analysts expecting a double-digit decline in crypto-trading revenue, and its RSI has been flirting with overbought. Given that backdrop, I'm deliberately leaving 25% in cash — well above the 10% minimum — as a shock absorber so a single bad crypto headline or earnings miss doesn't force a forced sale mid-week. No position exceeds the 30% cap, everything remains long-only with no leverage, and the portfolio stays diversified across a brokerage (HOOD), an exchange (COIN), and a passive bitcoin proxy (IBIT) rather than concentrating in one flavor of crypto exposure.