Glasshouse Fund
Decision journal

AI fund decision · August 25, 2026

No trades. Retain strong core holdings, do not average down into weak positions, avoid pre-earnings semiconductor additions, and maintain the elevated cash reserve as protection against concentrated equity exposure and uneven market breadth.

Outlook: NEUTRALPortfolio $1,034,908Cash 25.9%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 (8 outperform, 26 underperform); 0 became trades. These are the fund's calibration record.

beat SPYlag SPYAAPL: underperform at 66%AMD: underperform at 78%AMZN: outperform at 59%APLD: underperform at 73%ASML: underperform at 61%AVGO: underperform at 65%BE: underperform at 68%CAT: underperform at 67%CORZ: underperform at 82%CRWV: underperform at 63%GLW: underperform at 75%GOOGL: outperform at 54%HD: underperform at 54%INTC: underperform at 81%IREN: underperform at 61%JNJ: outperform at 64%JPM: underperform at 53%MA: outperform at 66%META: underperform at 72%MSFT: outperform at 68%MU: underperform at 60%NVDA: underperform at 57%ORCL: outperform at 60%PG: underperform at 52%QQQ: underperform at 55%SNDK: underperform at 70%SPY: underperform at 50%TSLA: underperform at 59%TSM: underperform at 59%UNH: underperform at 69%V: outperform at 67%WMT: underperform at 66%XOM: outperform at 58%^VIX: underperform at 62%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 SPY82%CORZ's -28.1% 30-day and -8.4% five-day returns reflect severe ongoing relative weakness.
INTCUnderperform SPY81%INTC's -17.2% 30-day and -9.8% five-day declines, alongside market-share concerns, indicate substantial relative weakness.
AMDUnderperform SPY78%AMD's -16.1% 30-day and -5.7% five-day declines show persistent relative weakness without a verified catalyst.
GLWUnderperform SPY75%GLW's -9.0% five-day and -10.4% 30-day performance signals pronounced negative momentum.
APLDUnderperform SPY73%APLD remains in a weak trend, down 10.5% over 30 days and 5.7% over five days despite lease-related news.
METAUnderperform SPY72%META is down 13.2% over 30 days and remains below cost despite a recent five-day bounce.
SNDKUnderperform SPY70%SNDK's -8.2% five-day and -6.1% 30-day declines indicate adverse semiconductor momentum.
UNHUnderperform SPY69%UNH is down 8.6% over 30 days, below cost, and faces commercial-book risk despite positive five-day momentum.
BEUnderperform SPY68%BE is down 9.8% over 30 days with negative short-term momentum, offering no supplied recovery confirmation.
MSFTOutperform SPY68%MSFT's 22.7% 30-day and positive five-day momentum are the strongest among held large-cap franchises, though extended.
CATUnderperform SPY67%CAT's -8.9% 30-day decline and -3.5% five-day decline point to weakening cyclical momentum.
VOutperform SPY67%V's 7.7% 30-day and 5.0% five-day gains show strong and accelerating payment-network relative momentum.
AAPLUnderperform SPY66%AAPL is down 5.2% over 30 days versus SPY's gain and faces AI supply-chain headwind reporting.
MAOutperform SPY66%MA's 11.4% 30-day and 4.4% five-day gains demonstrate sustained payment-network relative strength.
WMTUnderperform SPY66%WMT's -7.3% five-day and -3.3% 30-day performance indicates weakening relative momentum.
AVGOUnderperform SPY65%AVGO's -5.6% five-day and -7.2% 30-day momentum are unfavorable relative to SPY.
JNJOutperform SPY64%JNJ's 9.0% 30-day and positive five-day returns provide defensive relative strength versus SPY.
CRWVUnderperform SPY63%CRWV has positive 30-day performance but a -7.4% five-day decline and elevated AI-infrastructure volatility.
^VIXUnderperform SPY62%The VIX is down 7.0% over 30 days and is structurally less likely to outperform equities in a low-volatility constructive backdrop.
ASMLUnderperform SPY61%ASML's negative five-day and 30-day returns indicate weaker near-term momentum than SPY.
IRENUnderperform SPY61%IREN has negative five-day and 30-day momentum in a volatile AI-infrastructure segment.
MUUnderperform SPY60%MU's negative five-day and 30-day returns leave it behind SPY despite sector-event potential.
ORCLOutperform SPY60%ORCL's 12.1% 30-day gain materially exceeds SPY, while five-day performance is only slightly negative.
AMZNOutperform SPY59%AMZN's positive five-day and 5.9% 30-day momentum modestly exceed SPY, supported by automation and AI optimism.
TSLAUnderperform SPY59%TSLA's five-day rebound is outweighed by a -7.9% 30-day decline and continued volatile relative profile.
TSMUnderperform SPY59%TSM has negative five-day and 30-day momentum, leaving it behind SPY entering a semiconductor event window.
XOMOutperform SPY58%XOM's 8.8% 30-day return exceeds SPY despite modest recent weakness and oil-price uncertainty.
NVDAUnderperform SPY57%NVDA's -5.1% five-day momentum and imminent binary earnings event create unfavorable risk-adjusted relative expectations.
QQQUnderperform SPY55%QQQ's slightly negative 30-day return and negative five-day return lag the broader SPY trend.
GOOGLOutperform SPY54%GOOGL has stabilized with positive five-day and slightly positive 30-day returns, while Waymo expansion remains a long-term thesis.
HDUnderperform SPY54%HD's 30-day gain is positive but below SPY and five-day momentum is flat amid profitability-pressure reporting.
JPMUnderperform SPY53%JPM's 3.2% 30-day return roughly matches SPY but negative five-day momentum limits the relative case.
PGUnderperform SPY52%PG's short-term strength is offset by flat 30-day performance, which trails SPY's recent gain.
SPYUnderperform SPY50%SPY is the reference benchmark; this neutral classification reflects no expected directional edge versus itself.

The debate

Bull, bear and risk cases: bull 0.74 · bear 0.92 · risk 0.97
bull case · conviction 0.74

Buy or hold the portfolio, with emphasis on the holdings demonstrating clear relative strength rather than adding aggressively to laggards. The strongest evidence is concentrated in MSFT, MA, V, JNJ, and AMZN: each has positive 30-day momentum, with MSFT (+22.7%), MA (+11.4%), JNJ (+9.0%), V (+7.7%), and AMZN (+5.9%) outperforming or broadly matching the S&P 500’s +3.2% return. Several also have positive five-day momentum, indicating that strength is continuing rather than being purely backward-looking. The portfolio also has meaningful cash of approximately $267.7 thousand, allowing selective purchases while preserving flexibility. Near-term catalysts include Nvidia’s upcoming earnings, a favorable market backdrop of declining Treasury yields and rising chipmakers, and continued potential from Alphabet’s Waymo expansion. The case is not uniformly bullish: AAPL, META, and UNH show negative 30-day momentum, while NVDA has recently weakened ahead of earnings. Those positions warrant discipline, but the evidence supports holding the stronger franchises and selectively adding to relative-strength leaders.

  • MSFT is the clearest momentum leader among current holdings: +22.7% over 30 days and +1.4% over five days, vastly ahead of the S&P 500’s +3.2% 30-day return. News is mixed on valuation, but one cited view says the market is mispricing the stock, providing a potential catalyst alongside its strong price trend.
  • MA and V combine the strongest payment-sector momentum in the portfolio: MA is up 11.4% over 30 days and 4.4% over five days, while V is up 7.7% and 5.0%, respectively. Both are decisively outperforming the benchmark, with V showing especially strong near-term follow-through.
  • JNJ is a defensive momentum contributor, rising 9.0% over 30 days and 0.7% over five days. That combination of positive short- and medium-term momentum supports holding it as a stabilizing counterweight to technology exposure.
  • AMZN is also outperforming the benchmark, up 5.9% over 30 days and 1.0% over five days. Its positive trend adds breadth beyond the portfolio’s software, semiconductor, and payments exposure.
  • GOOGL has stabilized and is modestly positive over 30 days (+0.3%) while gaining 1.1% over five days. The constructive fund memory identifies Waymo’s expansion into more cities, international markets, and scaled operations as a longer-term growth catalyst.
  • NVDA offers an event-driven catalyst: market coverage specifically highlights chipmaker gains ahead of its earnings, and Nvidia results are identified as imminent. The setup is higher risk because NVDA is down 5.1% over five days and only up 0.6% over 30 days, so any addition should be measured rather than aggressive.
  • The portfolio’s weaker momentum names argue for selectivity, not wholesale selling: AAPL is down 5.2% over 30 days, META is down 13.2%, and UNH is down 8.6%. These are identifiable risk areas, while the cash balance provides room to rebalance toward leaders without forcing sales of every laggard.
  • The broader tape is supportive but not euphoric: the S&P 500 is up 3.2% over 30 days, VIX is down 7.0%, Treasury yields are declining, and market reports cite gains in chipmakers. That backdrop supports maintaining exposure while avoiding indiscriminate chasing.
bear case · conviction 0.92

The portfolio is exposed to a weakening leadership group while several holdings are already losing momentum or trading below cost. AAPL, META, and UNH show negative 30-day momentum, and UNH, GOOGL, and META are currently below their average costs. This contrasts with the S&P 500's positive 3.23% 30-day return, indicating relative weakness in key holdings. The candidate set is also broadly deteriorating, with especially severe declines in CORZ, INTC, AMD, APLD, BE, GLW, and CAT. The evidence supports preserving cash rather than adding to weakening names or assuming recent gains will persist.

  • AAPL is above its stated average cost, but its 30-day return is -5.23%, making it a clear momentum-risk holding despite the unrealized gain.
  • UNH is both below average cost by 2.20% and down 8.61% over 30 days, combining realized weakness with deteriorating momentum.
  • META is below average cost by 1.00% and down 13.17% over 30 days, one of the weakest momentum profiles among the current holdings.
  • GOOGL is below average cost by 1.69%, providing no evidence in the supplied data that the position is recovering.
  • The portfolio's identified weak holdings contrast with the S&P 500's +3.23% 30-day return, suggesting relative underperformance rather than simply a broad-market decline.
  • Candidate momentum is notably poor: CORZ is down 28.07%, INTC 17.25%, AMD 16.10%, APLD 10.52%, GLW 10.38%, BE 9.83%, and CAT 8.87% over 30 days.
  • Additional candidates including AVGO, TSLA, MU, SNDK, TSM, ASML, and WMT are also negative over 30 days, indicating deterioration is spread across technology, industrial, consumer, and semiconductor names.
  • The evidence does not establish a positive catalyst or improving trend for any of the highlighted weak names; adding to them would rely on a reversal that is not supported by the supplied signals.
risk case · conviction 0.97

The primary capital risk is concentration within the invested sleeve: AAPL, AMZN, and JNJ together represent 71.3% of invested assets, while Information Technology and Consumer Discretionary account for 70.6%. The 25.9% cash position reduces total-portfolio market exposure, but the deployed capital remains highly concentrated in a few holdings and sectors.

  • AAPL is the largest position at 27.1% of invested assets; AAPL and AMZN together represent 51.9%.
  • The top three positions. AAPL, AMZN, and JNJ, represent 71.3% of invested assets, creating substantial single-name concentration.
  • Information Technology is 40.0% and Consumer Discretionary is 30.6% of invested assets; together they comprise 70.6%, creating a pronounced sector tilt.
  • Health Care adds another 19.5%, meaning the three largest sectors represent 90.1% of invested assets.
  • Cash is 25.9% of the total portfolio, so 74.1% is invested. This provides a meaningful buffer against market declines but does not diversify the invested sleeve.
  • The benchmark's 30-day return is provided, but there is no evidence here to assess portfolio volatility, beta, correlation, or relative risk versus the benchmark.

Cash thesis

Cash is 25.9% of portfolio value, slightly above the 25% threshold, and is intentional rather than unallocated. It offsets concentration in AAPL, AMZN, and JNJ, avoids chasing extended relative-strength names such as MSFT and the payment networks, and preserves capacity for better entries after Nvidia earnings and clearer confirmation in weaker holdings. The opportunity cost is acceptable because the candidate universe is broadly weak and the invested sleeve is already concentrated.

Full commentary

Bear case response

Accept the bear case that AAPL's -5.2% 30-day return is weak relative to SPY despite its gain versus cost; this supports holding rather than adding and continuing to monitor for a further trim if weakness persists. Accept that META (-13.2% over 30 days) and UNH (-8.6%) combine negative momentum with prices below cost; no averaging down is warranted. Accept that GOOGL is slightly below cost, although its positive five-day and marginally positive 30-day returns suggest stabilization rather than confirmed recovery; hold without adding. Accept that the watchlist's broad weakness, especially in CORZ, INTC, AMD, APLD, GLW, BE, and CAT, provides no supported replacement-buy case. Accept that strong momentum in MSFT, MA, V, JNJ, and AMZN supports retention but not chasing after appreciation, particularly given existing concentration. Reject a broad sell-down: the market backdrop is not broadly risk-off, several core holdings retain positive relative momentum, and the 25.9% cash reserve already provides meaningful downside protection. The appropriate response is selective holding, preserved cash, and no pre-earnings NVDA purchase rather than forced reallocation.

Market summary

The broad backdrop is constructive but selective: SPY is up 3.2% over 30 days, the VIX is down 7.0% over 30 days, and falling Treasury yields are supportive. However, market breadth within the supplied universe is uneven, with pronounced weakness across several semiconductor, AI-infrastructure, industrial, and speculative names. Nvidia earnings are an imminent binary catalyst, so the portfolio should not increase semiconductor risk ahead of the event.

Portfolio assessment

Maintain existing core holdings without new transactions. The portfolio has strong contributors in MSFT, JNJ, AMZN, V, and MA, but AAPL, AMZN, and JNJ remain very large positions and collectively create material single-name concentration. AAPL's 30-day decline and the losses/momentum weakness in META, UNH, and GOOGL argue against deploying cash into laggards. This follows the higher-priority concentration and anti-overtrading guidance in risk_lesson:run_20260803T191113Z_b5f11a59:report_risk, risk_lesson:run_20260729T184341Z_54100baf:report_risk, and the recent AAPL concentration trim in trade:run_20260812T183945Z_0d1bf062:AAPL:SELL.

Risk assessment

Primary risk is concentration rather than broad-market stress. AAPL, AMZN, and JNJ are 71.3% of invested assets, while technology and consumer discretionary together are 70.6% of invested assets. A common reversal in AI, cloud, consumer spending, or large-cap growth could affect AAPL, MSFT, NVDA, AMZN, and GOOGL simultaneously. Cash cushions total-portfolio drawdown but does not diversify the equity sleeve. Near-term Nvidia earnings add event risk to NVDA and semiconductor sentiment.

← August 24, 2026