Glasshouse Fund
Decision journal

AI fund decision · September 8, 2026

Hold all positions and retain the 27.7% cash reserve. Strong holdings remain intact, but weak breadth, concentration, and persistent underperformance in GOOGL, HD, and AMZN do not justify deploying cash or increasing turnover. Sources used include current 5-day and 30-day returns, market and symbol news, benchmark data, and fund memory IDs risk_lesson:run_20260729T184341Z_54100baf:report_risk, risk_lesson:run_20260803T191113Z_b5f11a59:report_risk, report_summary:run_20260826T160411Z_f30a1737, trade:run_20260812T183945Z_0d1bf062:AAPL:SELL, and trade:run_20260731T185916Z_e4b0cfaf:NVDA:SELL.

Outlook: NEUTRALPortfolio $1,034,429Cash 27.7%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 (21 outperform, 13 underperform); 0 became trades. These are the fund's calibration record.

beat SPYlag SPYAAPL: outperform at 58%AMD: underperform at 68%AMZN: underperform at 63%APLD: underperform at 66%ASML: outperform at 56%AVGO: underperform at 69%BE: outperform at 64%CAT: underperform at 62%CORZ: underperform at 67%CRWV: outperform at 57%GLW: underperform at 59%GOOGL: underperform at 72%HD: underperform at 66%INTC: outperform at 58%IREN: outperform at 65%JNJ: outperform at 72%JPM: outperform at 57%MA: outperform at 56%META: outperform at 68%MSFT: outperform at 59%MU: outperform at 67%NVDA: outperform at 69%ORCL: outperform at 66%PG: outperform at 52%QQQ: underperform at 54%SNDK: outperform at 70%SPY: outperform at 50%TSLA: outperform at 55%TSM: outperform at 61%UNH: underperform at 58%V: outperform at 58%WMT: underperform at 57%XOM: outperform at 60%^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
GOOGLUnderperform SPY72%The 10.3% 30-day decline, below-cost position, and lack of confirmed reversal evidence materially lag SPY.
JNJOutperform SPY72%Its 8.5% 30-day and 1.5% five-day gains make it a clear defensive momentum leader versus SPY.
SNDKOutperform SPY70%Very strong five-day and 30-day momentum favors outperformance, though the magnitude of gains creates substantial reversal risk.
AVGOUnderperform SPY69%The 14.4% 30-day decline and negative five-day return indicate persistent relative weakness.
NVDAOutperform SPY69%Positive 8.7% 30-day and 5.9% five-day returns demonstrate strong relative momentum, despite AI-theme reversal risk.
AMDUnderperform SPY68%The 7.9% 30-day decline despite a recent bounce indicates weak intermediate relative momentum.
METAOutperform SPY68%Strong positive five-day and 30-day momentum supports relative leadership versus SPY.
CORZUnderperform SPY67%A 21.6% 30-day decline dominates the recent rebound and leaves the trend materially weaker than SPY.
MUOutperform SPY67%Strong five-day and 30-day momentum reflect AI-memory leadership, though valuation and geopolitical sensitivity raise volatility risk.
APLDUnderperform SPY66%The 15.7% 30-day loss signals elevated volatility and damaged trend despite a strong five-day rally and lease-related news.
HDUnderperform SPY66%A 7.1% 30-day loss and below-cost position outweigh its small five-day recovery.
ORCLOutperform SPY66%Strong positive five-day and 30-day returns support continued relative strength, though the recent advance may be extended.
IRENOutperform SPY65%Positive 30-day and very strong five-day momentum, alongside AI-contract narrative support, favor outperformance despite high volatility.
BEOutperform SPY64%Strong 30-day and five-day momentum, supported by S&P 500 inclusion news, favors relative strength but with elevated event-driven volatility.
AMZNUnderperform SPY63%A 6.8% 30-day decline materially trails SPY despite a positive five-day rebound and profitability versus cost.
CATUnderperform SPY62%Its 7.1% 30-day decline is substantially worse than SPY despite a recent five-day rebound.
TSMOutperform SPY61%Positive five-day and 30-day momentum supports relative strength within semiconductors.
XOMOutperform SPY60%Positive 30-day performance and oil-price strength support relative performance, despite recent five-day weakness.
GLWUnderperform SPY59%Negative 30-day momentum and disclosed customer, demand, and execution risks outweigh the recent rebound.
MSFTOutperform SPY59%Positive 30-day momentum and constructive recommendation coverage support a modest relative-strength view despite consolidation.
AAPLOutperform SPY58%Positive 30-day momentum and product-event catalysts support modest relative strength versus SPY, tempered by event and supply-chain risk.
INTCOutperform SPY58%A strong five-day advance and less severe 30-day decline than several chip peers support a cautious relative-strength view.
UNHUnderperform SPY58%Negative 30-day momentum and a price slightly below cost provide insufficient evidence of recovery.
VOutperform SPY58%Positive five-day and 30-day momentum supports modestly better performance than SPY.
CRWVOutperform SPY57%Positive five-day momentum and a relatively contained 30-day decline support a modest relative-strength view, with high volatility.
JPMOutperform SPY57%Positive five-day and slightly positive 30-day performance provide modest relative-strength support.
WMTUnderperform SPY57%The 3.7% 30-day decline is weaker than SPY despite a positive five-day move.
ASMLOutperform SPY56%Flat 30-day performance and positive five-day momentum are modestly better than SPY, but conviction is limited.
MAOutperform SPY56%Positive 30-day momentum and durable profitability versus cost support modest outperformance despite a slightly negative five-day return.
TSLAOutperform SPY55%Positive 30-day momentum is constructive, but a flat-to-negative five-day return limits confidence.
^VIXUnderperform SPY55%The 30-day decline in implied volatility suggests lower volatility index levels relative to SPY, though geopolitical headlines can reverse this quickly.
QQQUnderperform SPY54%Its modest 30-day decline and narrow technology breadth suggest a slight lag versus SPY.
PGOutperform SPY52%Near-flat but slightly positive five-day momentum and defensive characteristics provide only a marginal edge over SPY.
SPYOutperform SPY50%SPY is the required comparison benchmark, so no directional edge relative to itself is asserted.

The debate

Bull, bear and risk cases: bull 0.72 · bear 0.88 · risk 0.97
bull case · conviction 0.72

Maintain a bullish, selective-hold posture and deploy capital toward the strongest relative-strength holdings rather than broad, indiscriminate buying. The portfolio has several names materially outperforming a benchmark that declined 0.45% over the last 30 days, while volatility has eased as the VIX fell 7.3% over 30 days. The strongest evidence supports continued exposure to JNJ, NVDA, META, AAPL, MSFT, and the payments names. Near-term weakness in the broader market, driven by rising oil prices and geopolitical concerns, creates volatility but has not invalidated the relative-strength leadership visible in key holdings.

  • JNJ is a clear defensive momentum leader: up 8.5% over 30 days and 1.5% over five days versus the S&P 500's 30-day decline; current price of $270.21 is also above the $239.58 average cost. Recent coverage includes a call to buy the dip, reinforcing a constructive catalyst narrative.
  • NVDA combines strong momentum and profitable positioning: up 8.7% over 30 days and 5.9% over five days, while the position is up from a $203.30 average cost to $225.86. The broader AI and memory-stock theme remains prominent in current market coverage, supporting continued exposure despite short-term volatility.
  • META is showing one of the strongest combinations among existing holdings, gaining 4.9% over 30 days and 6.6% over five days. The position is already profitable, with the current price of $616.45 versus a $568.43 average cost, and the positive short-term acceleration supports holding rather than reducing.
  • AAPL has positive 30-day momentum of 3.5% and trades above the $298.69 average cost at $316.18. The upcoming iPhone reveal and potential foldable iPhone provide identifiable catalysts, although the evidence also includes warnings about supply-chain risk and the possibility that a major launch could disappoint; this supports holding and selective buying rather than an aggressive increase.
  • MSFT remains above its average cost by a wide margin, at $492.61 versus $412.76, and has positive 30-day momentum of 1.6%. Recent coverage includes a new buy recommendation and commentary that the market may be mispricing the stock, providing a constructive catalyst while the shares consolidate over the last five days.
  • The payments holdings provide steady relative strength: V is up 1.7% over 30 days and MA is up 1.4%, both outperforming the benchmark over the same period. Their current prices are substantially above average costs, supporting continued exposure to profitable leaders even though MA was slightly down over five days.
  • The portfolio is not dependent on one narrow winner: JNJ, NVDA, META, AAPL, MSFT, JPM, V, and MA are all above their stated average costs. JPM is also modestly positive over 30 days and has gained 1.0% over five days, adding financial-sector diversification to the technology and healthcare leadership.
  • There is meaningful cash available, approximately $287,000, so the committee can buy selectively during market weakness rather than being forced to chase strength. The preferred approach is to add incrementally to demonstrated leaders while preserving flexibility.
  • The bullish case is selective, not universal. AMZN, GOOGL, HD, and UNH have negative 30-day momentum, and current market headlines describe a sharp risk-off session tied to oil and geopolitical developments. Those names warrant patience and position-specific review rather than indiscriminate averaging down.
bear case · conviction 0.88

The clearest downside case is concentrated in HD, GOOGL, AMZN, and UNH: each is underperforming recently, while GOOGL, HD, and UNH are already below cost. The weakness is materially worse than the S&P 500's modest 30-day decline of 0.45%, suggesting deteriorating momentum rather than isolated portfolio noise. Candidate breadth is also unfavorable, with several technology, infrastructure, industrial, and consumer names posting materially negative 30-day returns, raising the risk of adding into weakening trends.

  • GOOGL is the most concerning current holding: it is down 3.83% versus average cost, while its 30-day return is down 10.32%, more than twice the decline of the S&P 500.
  • HD is down 3.79% versus cost and down 7.13% over 30 days, indicating both existing loss and continued deterioration.
  • AMZN remains above average cost, but its 30-day return is down 6.82%, showing meaningful momentum damage despite the unrealized gain.
  • UNH is marginally below cost and down 2.55% over 30 days, underperforming the benchmark and providing little evidence of near-term recovery.
  • The candidate set contains substantial downside momentum: CORZ is down 21.57%, APLD 15.67%, AVGO 14.41%, AMD 7.91%, CAT 7.14%, and WMT 3.72% over 30 days. These are not attractive entry signals from the evidence provided.
  • GLW is down 3.31% over 30 days, and the fund's prior risk disclosures explicitly identify customer losses, customer financing problems, product or component issues, technology change, litigation, regulatory action, and difficulty pacing capital spending to demand as risks.
  • Broad market confirmation is weak: the S&P 500 is down 0.45% over 30 days, while QQQ is down 0.68% and SPY is down 0.15%. The portfolio's identified weak names are therefore declining substantially more than broad benchmarks.
  • There is no downside evidence in the supplied data for AAPL, MSFT, NVDA, JPM, V, MA, or META; however, the evidence also does not establish that these holdings are immune to the broader deterioration or that new purchases in them are timely.
risk case · conviction 0.97

The main capital risk is concentration: 70.7% of invested assets are in AAPL, AMZN, and JNJ, while 70.3% of invested assets are concentrated in Information Technology, Consumer Discretionary, and Health Care. Cash at 27.7% provides a meaningful liquidity buffer and reduces total portfolio exposure, but it does not materially diversify the invested portion.

  • AAPL is the largest position at 28.5% of invested assets, creating the greatest single-name loss sensitivity.
  • AMZN and JNJ add another 42.2% of invested assets; the top three positions together represent 70.7%.
  • Information Technology is 42.1% of invested assets, led by AAPL at 28.5% and MSFT, NVDA, and GOOGL adding further technology exposure.
  • Consumer Discretionary is 28.2%, primarily driven by AMZN at 22.6% and HD at 5.6%.
  • Health Care is 19.8%, largely concentrated in JNJ at 19.6%; the sector is therefore almost a single-name exposure.
  • The portfolio has limited diversification outside the three largest sectors: Communication Services is 7.8% and Financials 2.2%.
  • Cash is 27.7% of the portfolio, which cushions portfolio-level drawdowns and provides liquidity, but the remaining 72.3% is still highly concentrated.
  • The benchmark's 30-day return is only evidence of a modest benchmark decline and does not establish that this portfolio's concentration risk is favorable or unfavorable relative to it.

Cash thesis

Cash is $286,974, or 27.7% of portfolio value. This reserve is intentional: it moderates total-portfolio drawdown from concentrated AAPL, AMZN, and JNJ exposures, preserves flexibility during an oil- and geopolitically driven risk-off session, and avoids forcing purchases into weak market breadth. The opportunity cost is acceptable until trend confirmation improves or a clearly diversified, high-conviction entry appears.

Full commentary

Bear case response

Accept: GOOGL is the clearest weak holding, down 10.3% over 30 days, below cost, and materially lagging SPY; do not average down without stabilization. Accept: HD is down 7.1% over 30 days and below cost, so no new capital should be committed. Accept: AMZN's 6.8% 30-day decline is meaningful despite its unrealized gain; it is a hold, not a buy. Accept: UNH is slightly below cost with negative 30-day momentum and lacks sufficient recovery evidence to add. Accept: weak 30-day returns in CORZ, APLD, AVGO, AMD, CAT, WMT, and GLW argue against broad watchlist deployment; GLW's disclosed operational and customer risks reinforce that restraint. Accept: modestly negative SPY and QQQ performance does not establish a broad bullish regime, and lower VIX is not proof of healthy breadth. Partially accept: the bear case supports trimming weak holdings eventually, but not an immediate sale today because HD and GOOGL have positive five-day returns, losses versus cost remain moderate, and a single risk-off session is insufficient confirmation for further turnover. Reject: the implication that all current leaders should be treated defensively; JNJ, NVDA, META, AAPL, MSFT, JPM, V, and MA have better momentum and should be retained, while no new purchases are made.

Market summary

The market is risk-off today, with headlines citing rising oil prices, geopolitical disruption, and declines in major indices. Over 30 days, SPY is nearly flat-to-down (-0.15%) and QQQ is modestly lower (-0.68%), while leadership is narrow: JNJ, NVDA, META, and selected semiconductor and infrastructure names are strong, but several cyclicals, consumer, AI-infrastructure, and technology names have sharply negative 30-day momentum. The low VIX level does not outweigh weak breadth or the current oil-driven macro uncertainty.

Portfolio assessment

The portfolio has appreciated to $1.034 million and retains profitable core exposure in JNJ, AAPL, MSFT, NVDA, AMZN, and payment networks. However, concentration remains the governing issue: AAPL, AMZN, and JNJ represent the majority of invested assets, while technology and consumer-discretionary exposures remain correlated. Existing positions in GOOGL and HD have negative 30-day momentum and are below cost; AMZN is still profitable but has weakened materially over 30 days. Consistent with recent low-turnover fund decisions, no trade is warranted today because neither broad confirmation nor a company-specific deterioration threshold supports action.

Risk assessment

The principal risk is concentration rather than liquidity. AAPL, AMZN, and JNJ dominate invested capital, so adverse consumer, technology, regulatory, or company-specific developments could have an outsized impact. AAPL and AMZN also create overlap between technology and consumer-discretionary risk, while JNJ largely drives healthcare exposure. GOOGL and HD show persistent negative 30-day relative performance; AMZN's decline warrants monitoring despite its gain versus cost. The cash reserve limits exposure but does not diversify the invested book. Recent fund memory requires avoiding repeated turnover and using selective, incremental risk reduction only when deterioration persists.

← September 7, 2026