Glasshouse Fund
Decision journal

AI fund decision · September 3, 2026

No trades. Maintain a neutral, low-turnover stance, retain the 25.4% cash reserve, hold proven leaders, and avoid adding to either concentrated winners or weak momentum names. Sources used include market returns, market and symbol news, and fund memory IDs risk_lesson:run_20260803T191113Z_b5f11a59:report_risk, risk_lesson:run_20260729T184341Z_54100baf:report_risk, trade:run_20260812T183945Z_0d1bf062:AAPL:SELL, and report_summary:run_20260826T160411Z_f30a1737.

Outlook: NEUTRALPortfolio $1,052,074Cash 25.4%Grounding: ok

Trades

No trades this day. The fund held.

Market calls

A directional call (beat or lag the S&P 500 over the horizon) on every researched name, whether or not the fund traded it. 34 calls (20 outperform, 14 underperform); 0 became trades. These are the fund's calibration record.

beat SPYlag SPYAAPL: underperform at 57%AMD: underperform at 65%AMZN: outperform at 62%APLD: underperform at 73%ASML: underperform at 55%AVGO: underperform at 61%BE: outperform at 56%CAT: underperform at 60%CORZ: underperform at 76%CRWV: underperform at 58%GLW: outperform at 58%GOOGL: underperform at 56%HD: underperform at 63%INTC: underperform at 59%IREN: outperform at 55%JNJ: outperform at 68%JPM: outperform at 57%MA: outperform at 58%META: outperform at 69%MSFT: outperform at 68%MU: outperform at 67%NVDA: outperform at 66%ORCL: outperform at 61%PG: outperform at 59%QQQ: outperform at 54%SNDK: outperform at 72%SPY: outperform at 50%TSLA: outperform at 62%TSM: outperform at 56%UNH: underperform at 60%V: outperform at 60%WMT: underperform at 54%XOM: outperform at 64%^VIX: underperform at 55%50%60%70%80%90%100%
Each tick is one call, placed by stated confidence. Hover for the name.
All 34 calls, with reasoning
SymbolCallConf.Why
CORZUnderperform SPY76%A 22.93% 30-day loss and continued five-day weakness outweigh data-center expansion headlines.
APLDUnderperform SPY73%The 10.94% 30-day and 9.22% five-day declines show an unresolved downtrend without a confirmed catalyst.
SNDKOutperform SPY72%Strong five-day and 30-day gains demonstrate exceptional momentum, albeit with elevated volatility risk.
METAOutperform SPY69%Positive five-day and 30-day momentum materially exceed SPY, although the position should not be chased.
JNJOutperform SPY68%Strong five-day and 30-day gains provide defensive relative strength in an uneven market.
MSFTOutperform SPY68%A 10.34% 30-day gain and Azure reporting catalyst support continued relative strength despite a modest five-day pullback.
MUOutperform SPY67%Positive returns over five and 30 days show unusually strong semiconductor relative momentum.
NVDAOutperform SPY66%Its 15.06% 30-day gain is the strongest portfolio momentum signal, though recent consolidation lowers confidence.
AMDUnderperform SPY65%Negative five-day and 30-day momentum indicate a weaker semiconductor setup than the broad market.
XOMOutperform SPY64%Positive oil-related market context and positive returns across both periods support relative strength.
HDUnderperform SPY63%It is below cost and down 4.45% over 30 days, with recent earnings-related price pressure.
AMZNOutperform SPY62%An 8.27% 30-day gain materially exceeds SPY despite a modest recent pullback, supporting relative-strength leadership.
TSLAOutperform SPY62%Strong 30-day gains and Cybercab-event momentum support relative strength, with event-driven volatility limiting conviction.
AVGOUnderperform SPY61%Negative returns over both five and 30 days leave its current relative trend below SPY.
ORCLOutperform SPY61%A 14.26% 30-day advance and upcoming earnings test support relative strength, tempered by a 4.07% five-day decline.
CATUnderperform SPY60%Negative five-day and 30-day returns indicate cyclical weakness versus SPY.
UNHUnderperform SPY60%Negative 30-day performance, below-cost status and governance/margin uncertainty outweigh the recent small rebound.
VOutperform SPY60%A positive 30-day return with minimal five-day weakness supports continued modest relative resilience.
INTCUnderperform SPY59%Negative five-day and 30-day returns indicate insufficient relative-strength confirmation.
PGOutperform SPY59%Positive returns across both periods provide defensive momentum versus a declining benchmark.
CRWVUnderperform SPY58%Positive 30-day momentum is offset by a sharp 6.76% five-day drop and highlighted financing risk.
GLWOutperform SPY58%The 6.59% 30-day gain remains favorable relative to SPY despite a weak recent week.
MAOutperform SPY58%Its positive 30-day return and strong portfolio-level execution support moderate relative strength.
AAPLUnderperform SPY57%The five-day rebound is constructive, but negative 30-day momentum and outsized portfolio concentration limit expected relative performance versus SPY.
JPMOutperform SPY57%Positive performance over both horizons and retreating yields support modest expected outperformance.
BEOutperform SPY56%Its positive 30-day return and relatively limited five-day decline suggest modest relative resilience.
GOOGLUnderperform SPY56%Near-flat 30-day performance and negative five-day momentum trail the stronger large-cap leaders, despite longer-term Waymo optionality.
TSMOutperform SPY56%Positive 30-day performance supports modest outperformance despite a weak five-day period.
ASMLUnderperform SPY55%Slightly positive 30-day performance is offset by a 3.04% five-day decline and weaker near-term semiconductor breadth.
IRENOutperform SPY55%Positive 30-day performance supports a modest relative-strength view, though near-term momentum is negative.
^VIXUnderperform SPY55%The VIX remains low after an 11.06% 30-day decline, although its recent rise argues against high conviction.
QQQOutperform SPY54%Its positive 30-day return versus SPY supports a narrow outperformance view, though recent technology weakness is a constraint.
WMTUnderperform SPY54%The strong five-day rebound does not fully offset negative 30-day momentum, leaving the signal mixed.
SPYOutperform SPY50%SPY is the comparison baseline; its direction is necessarily neutral relative to itself.

The debate

Bull, bear and risk cases: bull 0.80 · bear 0.86 · risk 0.96
bull case · conviction 0.80

Buy or hold the strongest leaders rather than de-risking into broad-market weakness. The S&P 500 declined 0.9% over 30 days, yet several portfolio holdings produced substantial positive momentum: NVDA rose 15.1%, MSFT 10.3%, META 10.0%, AMZN 8.3%, and JNJ 8.1%. This broad relative strength across technology, communications, consumer, and healthcare suggests the portfolio’s leading businesses are outperforming despite a soft benchmark. Near-term pullbacks in several names appear manageable against strong 30-day trends, while company-specific catalysts, including Microsoft’s planned standalone Azure revenue disclosure, Apple’s unusual multi-decade trading pattern and recent rally, and continued Waymo expansion within Alphabet, provide reasons to maintain exposure and selectively add to leaders.

  • Relative strength is compelling: NVDA (+15.1%), MSFT (+10.3%), META (+10.0%), AMZN (+8.3%), and JNJ (+8.1%) all materially outperformed the S&P 500’s -0.9% 30-day return.
  • NVDA is the strongest portfolio momentum name, up 15.1% over 30 days despite a modest -1.6% five-day pullback; that combination supports holding through normal consolidation rather than reacting to short-term weakness.
  • MSFT combines 10.3% 30-day momentum with a current catalyst: the company is preparing to disclose standalone Azure cloud revenues in a major reporting overhaul.
  • META has the best combination of recent and intermediate momentum among the portfolio: +3.8% over five days and +10.0% over 30 days.
  • AMZN is up 8.3% over 30 days, demonstrating strong relative strength even after a modest five-day decline.
  • JNJ is up 8.1% over 30 days and 3.6% over five days, offering a defensive source of momentum while the market remains volatile.
  • AAPL’s five-day gain of 3.3% and reported market behavior described as something it has not done in 20 years are constructive catalysts, although its 30-day return remains -2.5%.
  • Alphabet’s 30-day return is modestly positive at 1.0%, but the constructive thesis includes Waymo expanding to more domestic cities, entering international markets, and scaling operations.
  • The evidence is not uniformly bullish: HD (-4.5%), UNH (-5.2%), and AAPL (-2.5%) lagged over 30 days, and several leaders were down over five days. The strongest action is therefore selective buying or holding of relative-strength names, not indiscriminate buying.
bear case · conviction 0.86

The strongest bearish case is that several holdings and candidates are losing momentum while the benchmark is already down 0.9% over 30 days. AAPL, a significant holding by share count, has faded 2.5% over 30 days despite being above its average cost, creating risk that recent gains are reversing. HD, UNH, and GOOGL are both below cost and showing negative 30-day momentum, indicating deterioration rather than isolated mark-to-market weakness. The candidate list is even more vulnerable: APLD and CORZ have suffered severe 30-day declines, while AMD and AVGO are also materially weaker. Adding to these names would risk buying into established downward momentum.

  • AAPL has negative 30-day momentum of -2.46%, so its current gain versus average cost may be vulnerable to reversal.
  • HD is below average cost by approximately 2.64% and has declined 4.45% over 30 days, showing both realized portfolio weakness and deteriorating momentum.
  • UNH is below average cost by approximately 0.47% and has fallen 5.17% over 30 days, making the position's weakness more than a one-day fluctuation.
  • GOOGL is below average cost by approximately 2.55% and has negative recent performance, weakening the case for continued upside.
  • Candidate momentum is notably poor: APLD is down 10.94% over 30 days and CORZ is down 22.93%, while AMD and AVGO are down 5.84% and 5.31%, respectively.
  • CAT, INTC, and WMT also show negative 30-day momentum, suggesting weakness is not limited to one narrow candidate.
  • The S&P 500 is down 0.90% over the same 30-day period, so the portfolio is operating against a weakening market backdrop rather than a clearly supportive one.
  • The evidence does not justify adding to the weakest names until momentum stabilizes; the largest downside risk is compounding existing weakness by averaging into falling securities.
risk case · conviction 0.96

The largest capital risk is concentration: 71.4% of invested assets are in just AAPL, AMZN, and JNJ, while Information Technology and Consumer Discretionary together represent 71.0% of invested assets. Cash is a meaningful 25.4% of the portfolio and reduces total portfolio exposure, but the invested sleeve remains materially dependent on a small number of holdings and two sectors.

  • AAPL is 28.1% of invested assets, the single largest position; AMZN is 24.1% and JNJ is 19.2%. These three positions total 71.4% of invested assets.
  • The top two positions, AAPL and AMZN, account for 52.2% of invested assets, creating substantial single-name concentration.
  • Information Technology is 41.5% and Consumer Discretionary is 29.5% of invested assets; their combined 71.0% represents a strong sector tilt.
  • Health Care is also concentrated at 19.4%, largely alongside JNJ's 19.2% position, so diversification within that sector is limited by the evidence shown.
  • Cash is 25.4% of the total portfolio, which provides a capital buffer and lowers overall invested exposure, but does not diversify the risk of the invested sleeve.
  • The remaining positions are individually small, with six holdings each below 2.1% of invested assets, limiting their ability to offset concentration in the largest positions.
  • The benchmark's 30-day return of -0.9% does not provide enough evidence to assess relative portfolio performance or directional risk.

Cash thesis

Cash is 25.4% of portfolio value, slightly above the required threshold. The reserve is intentional: it offsets concentrated AAPL, AMZN and JNJ exposure, reduces participation in a potentially uneven market decline, and preserves capacity to buy only after broader confirmation or a company-specific setup with improving momentum. It is not a directional bet on a crash; it is risk budget discipline while leaders are extended and many watchlist names remain weak.

Full commentary

Bear case response

Accept: AAPL's -2.46% 30-day return warrants no new purchase despite its +3.30% five-day rebound; the existing concentration makes confirmation more important than chasing a rebound. Accept: HD, UNH and GOOGL are below cost with negative 30-day momentum, so they should be held without averaging down; their weakness has not yet met a threshold requiring a forced sale because positions are modest except GOOGL and no thesis-breaking fact is supplied. Accept: APLD, CORZ, AMD and AVGO have weak 30-day trends, so no purchases are appropriate before stabilization. Accept: the benchmark's -0.90% 30-day result and uneven breadth argue against broad deployment of cash. Partly reject: strong relative performance in NVDA, MSFT, META, AMZN and JNJ is sufficient to retain existing positions rather than de-risk indiscriminately, but not sufficient to add after material recent advances. The response follows higher-priority fund lessons to avoid repeated turnover, preserve liquidity, and use incremental trimming only if deterioration persists.

Market summary

The S&P 500 is down 0.9% over 30 days, while volatility remains comparatively subdued at 14.53 despite a 4.8% five-day VIX rise. Market breadth is uneven: AI/cloud leaders including NVDA, MSFT, META and AMZN retain strong 30-day performance, but several semiconductor, infrastructure and cyclical candidates are in active drawdowns. Falling Treasury yields are supportive, but upcoming jobs data and Gulf-related oil risk justify patience rather than broad risk expansion.

Portfolio assessment

The portfolio has performed adequately through a concentrated group of large winners, especially JNJ, AAPL, AMZN and MSFT. However, AAPL, AMZN and JNJ remain the dominant exposures, while AAPL and AMZN together represent a large share of total capital even after the existing cash reserve. Recent memory supports low turnover: AAPL was already trimmed on August 12 to reduce concentration, and the fund has repeatedly identified concentration and technology/consumer-discretionary correlation as its principal risk. Hold the core positions but do not add to them after recent gains or while market breadth is mixed.

Risk assessment

The main risk is concentrated, correlated equity exposure rather than insufficient liquidity. AAPL, AMZN and JNJ comprise roughly 50.4% of total portfolio value and 71.4% of invested assets. Technology and consumer-discretionary exposure can decline together if AI/cloud spending, consumer demand, regulation or valuation multiples weaken. HD, UNH and GOOGL are below cost with negative 30-day momentum, while strong gains in MSFT, NVDA, JNJ, V and MA carry mean-reversion risk. Cash mitigates total-capital drawdown but cannot fully offset losses in the largest holdings.

← September 2, 2026