Glasshouse Fund
Decision journal

AI fund decision · September 30, 2026

No executed buys or sells are warranted. Hold the concentrated core, retain the elevated cash reserve, and require improving breadth before deploying capital. Strong AI and semiconductor leadership supports maintaining existing exposure, but widespread weakness among AMZN, financials, healthcare, and several consumer holdings prevents a bullish allocation shift.

Outlook: NEUTRALPortfolio $1,045,389Cash 28.6%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 59%AMD: outperform at 57%AMZN: underperform at 63%APLD: underperform at 62%ASML: outperform at 61%AVGO: outperform at 52%BE: outperform at 58%CAT: outperform at 55%CORZ: underperform at 66%CRWV: underperform at 56%GLW: outperform at 58%GOOGL: underperform at 54%HD: underperform at 70%INTC: outperform at 56%IREN: underperform at 60%JNJ: underperform at 53%JPM: underperform at 64%MA: underperform at 60%META: outperform at 65%MSFT: outperform at 62%MU: outperform at 60%NVDA: outperform at 63%ORCL: underperform at 66%PG: outperform at 53%QQQ: outperform at 56%SNDK: outperform at 55%SPY: outperform at 50%TSLA: underperform at 61%TSM: outperform at 66%UNH: underperform at 60%V: underperform at 58%WMT: underperform at 60%XOM: outperform at 53%^VIX: outperform 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 SPY70%A 13.3% 30-day decline and negative five-day momentum are substantially weaker than SPY.
CORZUnderperform SPY66%Large negative five-day and 30-day returns indicate materially weaker momentum than SPY.
ORCLUnderperform SPY66%Negative five-day and 7.4% negative 30-day performance are materially weaker than SPY.
TSMOutperform SPY66%Positive five-day and 9.7% 30-day gains demonstrate durable semiconductor relative strength versus SPY.
METAOutperform SPY65%Its 28.3% 30-day return is exceptional relative strength, though the recent pullback and reversal risk limit confidence.
JPMUnderperform SPY64%Negative five-day and 6.0% negative 30-day momentum indicate weakening relative performance.
AMZNUnderperform SPY63%Negative returns over both five and 30 days versus a flat SPY indicate weaker relative momentum.
NVDAOutperform SPY63%Positive returns over both periods and the reported buyback-plan boost support relative strength, tempered by AI-expectation risk.
APLDUnderperform SPY62%The stock is down 6.7% over five days and 5.4% over 30 days, showing weak relative momentum.
MSFTOutperform SPY62%Positive five-day and 30-day momentum supports continued relative strength from cloud and AI exposure.
ASMLOutperform SPY61%Positive five-day and 30-day returns support semiconductor-equipment leadership versus SPY.
TSLAUnderperform SPY61%The 7.2% five-day decline and only modest 30-day gain indicate weak near-term relative momentum.
IRENUnderperform SPY60%The 12.1% five-day loss and mixed 30-day result indicate elevated volatility and weak near-term relative momentum.
MAUnderperform SPY60%A small five-day rebound does not offset the 5.8% 30-day decline versus SPY.
MUOutperform SPY60%A 13.5% 30-day gain supports semiconductor relative leadership despite a slightly negative five-day move.
UNHUnderperform SPY60%The 5.9% 30-day decline and position below cost indicate continuing relative weakness despite a positive five-day move.
WMTUnderperform SPY60%The 3.4% five-day decline and only modest 30-day gain show weaker near-term momentum than SPY.
AAPLOutperform SPY59%Positive 30-day relative strength and an October product-event catalyst support modest outperformance versus SPY, tempered by recent weakness and concentration.
BEOutperform SPY58%Strong five-day and 30-day gains show clear relative strength, although the sharp run raises volatility risk.
GLWOutperform SPY58%Positive returns over both horizons support relative strength, though filing-based execution and demand risks constrain conviction.
VUnderperform SPY58%Positive five-day momentum is outweighed by a 4.7% 30-day decline, leaving relative strength unconfirmed.
AMDOutperform SPY57%Its 26.3% 30-day gain indicates strong semiconductor relative strength, though the elevated run reduces confidence.
^VIXOutperform SPY57%Positive five-day and 30-day VIX moves indicate rising volatility demand, though this is a risk signal rather than an equity allocation recommendation.
CRWVUnderperform SPY56%Negative five-day and 30-day returns imply modest relative underperformance.
INTCOutperform SPY56%The 31.4% 30-day advance suggests strong relative strength despite a sharp recent pullback.
QQQOutperform SPY56%The 3.8% 30-day gain versus SPY's slight decline supports continued large-cap technology leadership.
CATOutperform SPY55%Positive short- and medium-term momentum modestly supports outperformance versus a flat benchmark.
SNDKOutperform SPY55%Strong 30-day performance supports outperformance, but a nearly 4.8% five-day decline signals high volatility.
GOOGLUnderperform SPY54%The near-flat but negative 30-day return and only modest five-day rebound provide insufficient evidence of sustained outperformance.
JNJUnderperform SPY53%Slightly negative five-day and 30-day returns suggest modest underperformance versus SPY despite its defensive characteristics.
PGOutperform SPY53%Positive five-day and 30-day returns modestly support defensive outperformance, although conviction is low.
XOMOutperform SPY53%Near-flat five-day performance and a positive 30-day gain provide a modest relative edge versus SPY.
AVGOOutperform SPY52%Near-flat recent returns are still slightly better than SPY's 30-day decline, but conviction is low because momentum is limited.
SPYOutperform SPY50%SPY is the comparison benchmark, so relative performance versus itself is neutral by definition.

The debate

Bull, bear and risk cases: bull 0.78 · bear 0.84 · risk 0.98
bull case · conviction 0.78

The strongest bullish case is to hold and selectively add to the portfolio's technology and semiconductor leaders rather than retreating to cash. Relative strength is concentrated in the companies most exposed to AI, cloud infrastructure, and advanced semiconductors: over 30 days, META gained 28.3%, TSM 9.7%, NVDA 8.5%, AAPL 5.1%, GLW 4.1%, and MSFT 2.5%, while the S&P 500 declined 0.2%. Several of these positions are also comfortably above cost, and the evidence points to tangible catalysts rather than momentum alone. The case is not universal. AMZN, GOOGL, financials, and several defensive holdings have weaker or negative 30-day momentum, while volatility is rising, but leadership remains strong enough to justify maintaining exposure and emphasizing the strongest names.

  • Clear relative strength versus the benchmark: the S&P 500 returned -0.2% over 30 days, versus +28.3% for META, +9.7% for TSM, +8.5% for NVDA, +5.1% for AAPL, +4.1% for GLW, and +2.5% for MSFT.
  • Momentum is broad across key AI and semiconductor holdings: NVDA is up 8.5% over 30 days, TSM 9.7%, and MSFT 2.5%; TSM also gained 2.3% over five days, while NVDA gained 0.8% and MSFT 1.7%.
  • META is the portfolio's standout momentum leader, up 28.3% over 30 days despite a modest -0.7% five-day pullback. That combination is consistent with profit-taking after substantial strength rather than evidence of a broken trend.
  • NVDA has a meaningful corporate catalyst: reported news says its share buyback plan received a $150 billion boost. Its positive 30-day momentum supports the view that investors continue to reward its AI positioning.
  • AAPL remains above cost and has positive 30-day momentum of 5.1%. The reported plan to unveil an all-new smart home hub at the October 13 event provides a near-term catalyst, even though the stock has pulled back 2.3% over five days.
  • MSFT combines positive five-day and 30-day momentum with reported investor attention around its AI opportunity. Its current price is materially above the stated average cost, supporting a hold-and-compound posture.
  • TSM is both above cost and one of the strongest large-cap holdings on momentum, with gains over both five and 30 days. Its relative strength provides confirmation that semiconductor exposure is not limited to a single company.
  • The portfolio has approximately 28.8% cash, according to the constructive fund memory, giving the committee flexibility to add selectively to proven leaders rather than needing to sell into weakness.
  • The bullish view should remain selective: AMZN is down 5.2% over 30 days, JPM is down 6.0%, MA down 5.8%, and volatility has risen, with the VIX up 5.5% over 30 days. Those facts argue for staged purchases and concentration in the strongest relative performers, not indiscriminate buying.
bear case · conviction 0.84

The strongest bearish case is that momentum is deteriorating across several major holdings and candidates despite an essentially flat S&P 500 over the last 30 days. Weakness is especially concentrated in consumer, healthcare, financial, and large-platform names, while several current holdings are already below cost. The portfolio risks further downside if recent weakness broadens from laggards into otherwise profitable positions.

  • Five current holdings are below average cost: UNH (-9.0%), WMT (-3.6%), XOM (-3.5%), GOOGL (-1.0%), and PG (-0.8%).
  • Momentum is negative for 12 listed holdings or candidates, including AMZN (-5.2%), JPM (-6.0%), MA (-5.8%), V (-4.7%), UNH (-5.9%), JNJ (-0.9%), and GOOGL (-0.3%).
  • The deterioration is not confined to one industry: healthcare weakness appears in UNH, financial/payment weakness in JPM, MA, and V, consumer weakness in AMZN and WMT, and platform weakness in GOOGL.
  • Several candidates show substantially worse momentum than the benchmark: HD is down 13.3% over 30 days, ORCL 7.4%, CORZ 6.4%, APLD 5.4%, and CRWV 2.4%, versus the S&P 500 down only 0.2%.
  • AMZN, JPM, MA, V, and UNH combine meaningful negative 30-day momentum with substantial portfolio exposure, making continued weakness more consequential than isolated declines in small positions.
  • GLW's cautionary filings explicitly identify risks including inability to pace capital spending to customer demand, margin-improvement failure, technology change, litigation, product or component problems, customer financing and receivables risk, loss of significant customers, and regulatory developments.
  • The portfolio contains many positions with positive gains from cost, but that creates downside exposure if weakening momentum causes profitable positions to reverse; the evidence does not establish that current gains are supported by improving momentum.
  • Conviction is moderate-high rather than absolute: the breadth of negative momentum is concerning, but the evidence supplied contains limited operating or valuation data and does not show weakness in every holding.
risk case · conviction 0.98

The biggest capital risk is concentration: 67.4% of invested assets is allocated to AAPL, JNJ, and AMZN, while the top five positions represent 84.1%. Sector exposure is also highly concentrated, with Information Technology, Health Care, and Consumer Discretionary totaling 85.1% of invested assets. Cash at 28.6% provides a substantial buffer, but the invested portion remains dependent on a small number of positions and sectors.

  • AAPL is the largest position at 30.3% of invested assets, creating the single greatest position-level source of capital risk.
  • JNJ and AMZN add 19.3% and 17.8%, respectively; together with AAPL, the three largest positions comprise 67.4% of invested assets.
  • The top five positions. AAPL, JNJ, AMZN, MSFT, and GOOGL, account for 84.1% of invested assets, leaving limited diversification among the invested holdings.
  • Information Technology is 47.9% of invested assets, with additional meaningful exposure to Health Care at 19.4% and Consumer Discretionary at 17.8%.
  • The three largest sectors total 85.1% of invested assets, creating substantial sector concentration independent of any directional forecast.
  • Cash is 28.6% of the portfolio, so 71.4% is invested; this reduces overall portfolio exposure but does not eliminate the concentration risk within invested capital.
  • The provided benchmark's 30-day return of approximately -0.2% does not materially change the structural concentration assessment.

Cash thesis

Cash is 28.6% of portfolio value, above the 25% threshold, and is intentional. It offsets concentrated exposure to AAPL, JNJ, AMZN, and AI-linked equities while market breadth is uneven and volatility is rising. The reserve also prevents forced selling and preserves capacity for selective deployment only after broader relative strength improves. Holding cash is preferable to averaging into weakening AMZN, GOOGL, financials, or defensives, or adding to already concentrated winners.

Full commentary

Bear case response

Accept the bear case on concentration: AAPL, JNJ, and AMZN dominate the invested sleeve, so new purchases in those names are not justified. Accept that weakness is broad across AMZN, JPM, MA, V, UNH, GOOGL, WMT, and several watchlist names; against a flat benchmark, their negative 30-day returns warrant a hold-and-monitor stance rather than averaging down. Accept that holdings below cost, especially UNH, WMT, XOM, GOOGL, and PG, require thesis discipline, but reject selling solely because they are below cost because the supplied evidence does not establish a fundamental thesis break. Accept that GLW's filing risks remain material, although its positive five-day and 30-day relative strength supports holding its modest position rather than treating those general risks as a sell trigger. Accept that gains in META, NVDA, AAPL, MSFT, and TSM can reverse and that catalysts do not prove improved fundamentals; therefore, hold rather than add. Reject an indiscriminate de-risking conclusion because the strongest leaders materially outperform SPY and the existing 28.6% cash allocation already provides meaningful downside moderation.

Market summary

The broad market is essentially flat over 30 days, with SPY down 0.2%, while leadership remains concentrated in AI, cloud, and semiconductor-linked names. TSM, NVDA, META, AAPL, MSFT, and GLW have outperformed, but several portfolio holdings across consumer, healthcare, payments, and financials have negative 30-day momentum. The VIX is up over both five and 30 days, supporting a selective rather than aggressive risk posture.

Portfolio assessment

The portfolio has attractive unrealized gains in AAPL, JNJ, MSFT, NVDA, META, TSM, and payment networks, but its main issue is structural concentration rather than insufficient equity exposure. AAPL, JNJ, and AMZN are very large holdings, while technology-linked positions remain correlated to AI spending and valuation risk. Maintain core winners without adding to the largest positions, monitor laggards for persistent deterioration, and avoid reactive turnover. This follows the low-turnover and concentration discipline in memory report_summary:run_20260929T193720Z_dac933f3 and risk_lesson:run_20260803T191113Z_b5f11a59.

Risk assessment

The principal risk is that 71.4% invested capital is concentrated in a small number of positions and sectors. AAPL is approximately 21.6% of total portfolio value, while JNJ and AMZN are also substantial. Technology, consumer discretionary, and healthcare exposures could decline together if AI-capex expectations, consumer demand, or rates deteriorate. Current company-specific risks include AMZN's negative 30-day momentum, GOOGL near/below cost, UNH weakness, and reversal risk in strong winners such as META, MSFT, NVDA, and TSM. Cash cushions total-portfolio volatility but does not eliminate concentrated-position risk.

← September 29, 2026 October 1, 2026 →