Glasshouse Fund
Decision journal

AI fund decision · September 22, 2026

Maintain a neutral, low-turnover stance: hold core winners, avoid adding to concentrated and extended technology positions, do not average down into laggards, and retain the 28.4% cash reserve for better-confirmed opportunities. Sources used include market momentum and news, risk_lesson:run_20260803T191113Z_b5f11a59:report_risk, risk_lesson:run_20260729T184341Z_54100baf:report_risk, report_summary:run_20260826T160411Z_f30a1737, and recent AAPL/NVDA trade-memory constraints.

Outlook: NEUTRALPortfolio $1,053,075Cash 28.4%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 (19 outperform, 15 underperform); 0 became trades. These are the fund's calibration record.

beat SPYlag SPYAAPL: outperform at 62%AMD: outperform at 58%AMZN: underperform at 56%APLD: outperform at 55%ASML: underperform at 59%AVGO: underperform at 55%BE: outperform at 57%CAT: outperform at 53%CORZ: outperform at 54%CRWV: underperform at 60%GLW: underperform at 56%GOOGL: outperform at 58%HD: underperform at 64%INTC: outperform at 60%IREN: outperform at 57%JNJ: underperform at 53%JPM: underperform at 57%MA: underperform at 55%META: outperform at 61%MSFT: outperform at 60%MU: outperform at 59%NVDA: outperform at 59%ORCL: outperform at 56%PG: underperform at 54%QQQ: outperform at 56%SNDK: outperform at 60%SPY: outperform at 50%TSLA: outperform at 55%TSM: outperform at 62%UNH: underperform at 62%V: underperform at 53%WMT: underperform at 59%XOM: underperform at 63%^VIX: underperform at 57%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
HDUnderperform SPY64%The 11.29% 30-day decline and negative five-day momentum indicate persistent weakness versus SPY.
XOMUnderperform SPY63%Negative five-day and 30-day performance, together with oil-price weakness, supports an underperform call.
AAPLOutperform SPY62%Positive 5-day and 30-day momentum, near-record-high coverage, and AI-server catalyst potential support relative strength, though concentration limits buy appeal.
TSMOutperform SPY62%Consistent 5-day and 30-day semiconductor leadership supports outperformance, tempered by geopolitical and cycle risk.
UNHUnderperform SPY62%Negative 30-day momentum and a loss versus portfolio cost indicate continuing relative weakness despite modest recent stabilization.
METAOutperform SPY61%Exceptional 30-day and five-day relative strength and a positive Muse-related catalyst support leadership, although extension risk is high.
CRWVUnderperform SPY60%The 8.31% 30-day decline is a material relative weakness signal despite a positive five-day move.
INTCOutperform SPY60%Very strong 5-day and 30-day gains alongside AI-spending optimism support continued relative strength, with elevated momentum risk.
MSFTOutperform SPY60%Positive 5-day and 30-day returns and constructive analyst commentary support continued moderate outperformance.
SNDKOutperform SPY60%Strong five-day and 30-day momentum and favorable AI-memory demand commentary support relative strength, with high extension risk.
ASMLUnderperform SPY59%The 30-day decline outweighs the recent bounce, leaving insufficient evidence of a durable relative-strength recovery.
MUOutperform SPY59%Strong memory-cycle momentum and supply-tightness commentary support relative strength, though the sharp advance increases cyclicality risk.
NVDAOutperform SPY59%Positive 5-day and 30-day momentum, AI demand, and potential China-policy upside support relative strength, but policy risk remains material.
WMTUnderperform SPY59%The 6.51% 30-day decline and negative five-day momentum point to continuing relative weakness.
AMDOutperform SPY58%Exceptional 5-day and 30-day semiconductor momentum supports continued relative strength, but the pace of the advance raises reversal risk.
GOOGLOutperform SPY58%Positive five-day and 30-day returns, plus longer-term Waymo expansion potential, support modest relative outperformance.
BEOutperform SPY57%Very strong 30-day performance and positive recent momentum support relative strength, although volatility and short-report risk remain high.
IRENOutperform SPY57%Positive momentum and reports of AI customer diversification support relative strength, though the business remains higher risk.
JPMUnderperform SPY57%Negative 30-day momentum and bank-sector volatility headlines weaken the relative outlook.
^VIXUnderperform SPY57%Declining 5-day and 30-day volatility trends are consistent with a lower-volatility environment and relative underperformance versus equities.
AMZNUnderperform SPY56%Slightly negative 30-day momentum and uncertainty around forward AWS-margin commentary weaken its near-term setup versus SPY.
GLWUnderperform SPY56%A sharp five-day rebound is not yet enough to offset flat-to-negative 30-day momentum and disclosed customer-demand and execution risks.
ORCLOutperform SPY56%Positive returns across both measured periods support a constructive cloud and AI-infrastructure relative trend.
QQQOutperform SPY56%Positive technology-sector momentum and AI leadership support modest outperformance versus the broader SPY benchmark.
APLDOutperform SPY55%Strong five-day acceleration can support near-term relative strength, but negative 30-day performance makes this a lower-confidence call.
AVGOUnderperform SPY55%Recent five-day strength is offset by negative 30-day momentum, suggesting an unconfirmed rebound rather than clear leadership.
MAUnderperform SPY55%Negative five-day and 30-day momentum suggest near-term relative underperformance despite the position's substantial gain versus cost.
TSLAOutperform SPY55%Positive five-day and 30-day momentum supports relative strength, though the name remains highly volatile.
CORZOutperform SPY54%Strong recent momentum supports upside participation, but the negative 30-day result and high-volatility profile reduce conviction.
PGUnderperform SPY54%Negative five-day momentum and only modest 30-day appreciation provide limited support for beating SPY.
CATOutperform SPY53%Positive five-day momentum supports a modestly constructive industrial outlook despite a negative 30-day return.
JNJUnderperform SPY53%Near-flat 30-day performance trails the benchmark and leaves limited evidence for near-term relative leadership despite positive absolute portfolio returns.
VUnderperform SPY53%Negative five-day momentum and only modest 30-day gains provide a less compelling near-term setup than SPY.
SPYOutperform SPY50%This is the benchmark itself; expected performance is effectively inline, so the required directional call carries neutral confidence.

The debate

Bull, bear and risk cases: bull 0.82 · bear 0.88 · risk 0.98
bull case · conviction 0.82

Buy or hold the portfolio’s strongest technology and AI-linked compounders: momentum and relative strength are decisively favorable in several core positions, while the market backdrop remains constructive. The S&P 500 gained 1.46% over 30 days, versus substantially stronger performance in AAPL (+9.34%), META (+36.45%), TSM (+7.96%), MSFT (+4.34%), NVDA (+3.59%), and GOOGL (+3.20%). Recent five-day acceleration in AAPL, META, NVDA, TSM, and GLW further supports an active leadership regime rather than a weakening trend. Catalysts include continued AI infrastructure demand, potential developments affecting Nvidia’s China access, Microsoft analyst optimism following a management meeting, Apple trading near record highs, Meta-related positive market reaction to Muse, and Alphabet’s Waymo expansion. The evidence supports adding selectively to leaders while maintaining cash for volatility; it does not support broad buying of every holding.

  • Relative strength is strongest in META, AAPL, TSM, MSFT, NVDA, and GOOGL: each materially outperformed the S&P 500’s 1.46% 30-day return.
  • META is the clearest momentum leader, gaining 36.45% over 30 days and 10.68% over five days; market news also reported that Muse added billions to Meta’s market capitalization.
  • AAPL gained 9.34% over 30 days and 2.31% over five days, with coverage describing shares as near record highs. Its portfolio position is also substantially above its average cost.
  • TSM delivered consistent momentum of 7.96% over 30 days and 7.88% over five days, providing exposure to the semiconductor cycle without relying on a single U.S. chip designer.
  • NVDA is up 3.59% over 30 days and 7.17% over five days. Reported catalysts include possible Trump–China discussions affecting AI chips, while coverage also notes that valuation has reached a decade low despite a profit boom.
  • MSFT’s 30-day gain of 4.34% beats the benchmark, and recent coverage cites a higher analyst target and additional upside following a management meeting.
  • GOOGL is showing positive acceleration, up 3.20% over 30 days and 2.90% over five days. The constructive fund evidence identifies Waymo’s expansion to more domestic cities, international entry, and operating scale as a longer-term catalyst.
  • The portfolio has approximately $299,000 in cash, allowing selective purchases of proven leaders rather than forcing immediate full deployment.
  • The evidence is mixed outside the leaders: AMZN, JNJ, PG, and V have limited or moderate momentum, while UNH, JPM, WMT, XOM, and MA have negative 30-day momentum. The bullish case is therefore concentrated, not indiscriminate.
bear case · conviction 0.88

The portfolio and candidate list show meaningful downside and weakening relative momentum despite the S&P 500 gaining 1.46% over the same 30-day period. The clearest risks are existing losses in UNH and XOM, fading momentum across several holdings, and materially worse trends among candidates such as HD, CRWV, ASML, and AVGO. This suggests a broad deterioration in selected names rather than isolated weakness, raising the risk of further drawdowns or adding to positions before momentum stabilizes.

  • UNH is already down 7.84% versus average cost and has negative 30-day momentum of 3.57%, making it the clearest combination of realized loss and continued deterioration.
  • XOM is down 5.82% versus average cost and has declined 4.39% over 30 days, indicating persistent weakness rather than a short-lived pullback.
  • AMZN, a current holding, has slightly negative 30-day momentum of 0.39%; combined with its limited gain versus average cost, this provides little evidence of a strong near-term trend.
  • WMT has fallen 6.51% over 30 days despite being only modestly above average cost, creating risk that the existing gain is eroded if weakness continues.
  • JPM, MA, JNJ, and GLW all have negative 30-day momentum, ranging from 0.11% to 3.09%, showing that weakness is spread across financial, healthcare, consumer, and materials-related holdings.
  • The candidate set contains substantially sharper declines: HD is down 11.29%, CRWV 8.31%, ASML 5.08%, and AVGO 4.39% over 30 days. These are poor entry conditions for adding exposure without evidence of stabilization.
  • GLW's disclosed risk factors include customer losses, customer financing problems, product or component performance issues, technology change, regulatory developments, litigation, and difficulty pacing capital spending to demand; its negative 30-day momentum provides no offsetting technical support.
  • The portfolio's weakening names are moving against a benchmark that gained 1.46% over 30 days, indicating potential relative underperformance and a risk that holding-specific problems are being masked by broader market strength.
risk case · conviction 0.98

The dominant capital risk is concentration: AAPL, JNJ, and AMZN comprise 67.9% of invested assets, while Information Technology alone represents 47.5%. Although the 28.4% cash allocation reduces total portfolio market exposure, the invested sleeve remains heavily dependent on a small number of positions and sectors.

  • AAPL is the largest position at 30.5% of invested assets, creating the greatest single-name exposure.
  • JNJ and AMZN add 19.4% and 18.0%, respectively; together with AAPL they account for 67.9% of invested assets.
  • Information Technology is 47.5% of invested assets, the largest sector allocation, with additional large exposures in Consumer Discretionary at 18.0% and Communication Services at 8.2%.
  • Health Care is also concentrated at 19.5%, largely represented by JNJ at 19.4% of invested assets.
  • The top five positions. AAPL, JNJ, AMZN, MSFT, and GOOGL, represent 84.1% of invested assets, leaving limited diversification within the invested portion.
  • Cash is 28.4% of the total portfolio, providing a substantial non-equity allocation, but it does not reduce concentration among the remaining invested assets.
  • The evidence does not show leverage or short exposure; the primary identified capital risks are single-name concentration and sector concentration rather than portfolio-wide market exposure.

Cash thesis

Cash is 28.4% of portfolio value, above the 25% threshold. This reserve is intentional: it limits total-capital exposure to the concentrated AAPL, JNJ, AMZN, and technology sleeve; preserves flexibility for confirmed pullbacks or diversification opportunities; and avoids forcing capital into momentum-chasing entries while several existing holdings underperform. The opportunity cost is meaningful in a rising market, but current dispersion and concentration make patience preferable to immediate deployment.

Full commentary

Bear case response

Accept: UNH and XOM combine losses versus cost with negative 30-day momentum, so they should not be averaged down without stabilization. Accept: AMZN has weak near-term relative momentum and its large size makes additional buying inappropriate. Accept: WMT, JPM, MA, JNJ, and GLW have lagged the benchmark over 30 days, supporting a hold-and-monitor stance rather than new allocations. Accept: HD, CRWV, ASML, and AVGO have negative 30-day trends and are not suitable new entries without confirmation. Accept: GLW has meaningful operational and customer-demand risks disclosed in fund memory, so its recent five-day rebound is insufficient to justify an add. Reject the implication of broad portfolio deterioration: AAPL, META, TSM, NVDA, MSFT, and GOOGL have outperformed SPY over 30 days, which supports retaining rather than broadly reducing quality technology leaders. Accept the rebuttal that momentum alone does not justify chasing these winners; their strong performance improves the hold case, not the buy case. Cash provides sufficient optionality, so there is no need to force either purchases into leaders or sales of core positions today.

Market summary

The market backdrop is constructive but narrow: SPY gained 1.46% over 30 days and VIX declined to 14.27, while AI and semiconductor leadership remains strong. AAPL, META, TSM, NVDA, MSFT, and GOOGL show favorable relative momentum, but several held defensive, financial, energy, and consumer names lag the benchmark. Geopolitical headlines involving Iran and China add event risk despite calm implied volatility.

Portfolio assessment

The portfolio has appreciated to $1.053 million, but concentration remains the primary issue. AAPL, JNJ, and AMZN comprise 67.9% of invested assets, while AAPL alone is roughly 21.8% of total portfolio value. Current leadership supports holding core technology exposure, but it does not justify adding to already-large positions after sharp advances. No trades are warranted today because the strongest names are extended, weaker names lack confirmed stabilization, and prior fund memory prioritizes low turnover and selective action.

Risk assessment

The dominant risk is not leverage or broad market stress, but correlated single-name and sector concentration. AAPL, JNJ, AMZN, MSFT, and GOOGL represent 84.1% of invested assets. Technology exposure remains substantial, and AI-linked holdings could decline together if spending expectations, policy conditions, or semiconductor demand weaken. UNH and XOM are below cost with negative 30-day momentum; WMT, JPM, MA, JNJ, and GLW also lagged over 30 days. Fund memory reinforces avoiding repeated turnover, maintaining liquidity, and using incremental trims only when deterioration is persistent rather than reacting to daily price movement.

← September 21, 2026 September 23, 2026 →