Glasshouse Fund
Decision journal

AI fund sells HD · September 17, 2026

Maintain a neutral, low-turnover posture. Trim HD once to address persistent deterioration, retain stronger core holdings without adding to concentrated leaders, avoid weak AI and semiconductor candidates, and preserve the substantial cash reserve until market breadth improves.

Outlook: NEUTRALPortfolio $1,043,159Cash 25.8%Grounding: ok

Trades

SELL 33 HD72% · filled at $303

Trim approximately one quarter of HD after a -10.83% 30-day decline and a -7.27% loss from cost. The position has no confirmed positive catalyst in the supplied evidence, and reducing it is consistent with the fund's higher-priority lesson to trim persistent weakening names incrementally rather than wait for a larger loss.

Risks the AI named

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 (11 outperform, 23 underperform); 0 became trades. These are the fund's calibration record.

beat SPYlag SPYAAPL: outperform at 70%AMD: underperform at 57%AMZN: underperform at 65%APLD: underperform at 75%ASML: underperform at 70%AVGO: underperform at 70%BE: outperform at 59%CAT: underperform at 61%CORZ: underperform at 69%CRWV: underperform at 76%GLW: underperform at 66%GOOGL: outperform at 56%HD: underperform at 71%INTC: underperform at 55%IREN: underperform at 60%JNJ: outperform at 64%JPM: underperform at 57%MA: outperform at 58%META: outperform at 73%MSFT: underperform at 54%MU: underperform at 63%NVDA: underperform at 59%ORCL: underperform at 68%PG: outperform at 60%QQQ: underperform at 62%SNDK: underperform at 69%SPY: outperform at 50%TSLA: outperform at 55%TSM: underperform at 58%UNH: underperform at 64%V: outperform at 61%WMT: underperform at 60%XOM: outperform at 53%^VIX: underperform at 53%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
CRWVUnderperform SPY76%The -6.47% 5-day and -21.58% 30-day declines outweigh backlog-related optimism and imply elevated downside risk versus SPY.
APLDUnderperform SPY75%A -20.35% 30-day decline and -5.39% 5-day decline make the AI-infrastructure candidate a weak risk-adjusted entry versus SPY.
METAOutperform SPY73%The strongest portfolio momentum profile, up 4.49% over 5 days and 13.17% over 30 days, supports continued outperformance versus SPY.
HDUnderperform SPY71%The -10.83% 30-day decline, loss versus cost and lack of a confirmed corrective catalyst point to continued weakness versus SPY.
AAPLOutperform SPY70%Strong 5-day and 30-day relative strength, plus iPhone promotion and product-cycle headlines, support continued outperformance versus SPY despite concentration risk.
ASMLUnderperform SPY70%The -5.05% 5-day and -13.30% 30-day returns show unresolved semiconductor-equipment weakness versus SPY.
AVGOUnderperform SPY70%AI exposure does not overcome sharply negative momentum of -5.91% over 5 days and -18.74% over 30 days.
CORZUnderperform SPY69%Despite favorable AI-pivot coverage, the -14.66% 30-day return indicates weak trend confirmation versus SPY.
SNDKUnderperform SPY69%A -10.20% 5-day decline reflects high volatility and inadequate trend support despite a nearly flat 30-day return.
ORCLUnderperform SPY68%The -6.39% 5-day and -8.36% 30-day returns indicate persistent relative weakness versus SPY.
GLWUnderperform SPY66%A -11.62% 5-day decline and -8.90% 30-day return indicate deteriorating momentum, with memory highlighting execution and customer-demand risks.
AMZNUnderperform SPY65%The -2.35% 5-day and -7.23% 30-day returns indicate persistent relative weakness versus SPY despite the AWS long-term narrative.
JNJOutperform SPY64%Positive 5-day and 30-day momentum, a 52-week-high reference and defensive characteristics support relative strength versus SPY.
UNHUnderperform SPY64%The position is below cost with negative 5-day and 30-day momentum, leaving insufficient evidence of a confirmed recovery versus SPY.
MUUnderperform SPY63%Negative 5-day and 30-day momentum in a weak semiconductor tape favor underperformance versus SPY.
QQQUnderperform SPY62%QQQ's -3.74% 30-day return trails SPY's -3.06%, reflecting technology-sector breadth weakness.
CATUnderperform SPY61%Negative 5-day and -8.41% 30-day momentum suggest cyclical industrial weakness versus SPY.
VOutperform SPY61%Positive 5-day and 30-day momentum supports continued defensive relative strength versus SPY.
IRENUnderperform SPY60%Negative 5-day and 30-day returns leave insufficient trend support for outperformance versus SPY.
PGOutperform SPY60%Positive 5-day and 30-day momentum and defensive demand characteristics support modest outperformance versus SPY.
WMTUnderperform SPY60%The 5-day rebound has not repaired the -6.88% 30-day decline or loss versus cost, favoring underperformance versus SPY.
BEOutperform SPY59%Positive data-center catalyst news and strong 5-day and 30-day momentum support relative strength, albeit with high volatility.
NVDAUnderperform SPY59%Constructive demand commentary is outweighed near term by negative 5-day and -4.95% 30-day momentum versus SPY.
MAOutperform SPY58%Stable positive 5-day and 30-day returns support modest outperformance versus SPY, though the signal is not strong.
TSMUnderperform SPY58%Negative 5-day and 30-day momentum amid semiconductor weakness modestly favor underperformance versus SPY.
AMDUnderperform SPY57%Positive short-term momentum is offset by a volatile semiconductor backdrop and weaker broad technology trend versus SPY.
JPMUnderperform SPY57%The -3.91% 30-day return and rate-related market uncertainty modestly favor underperformance versus SPY.
GOOGLOutperform SPY56%The 3.09% 5-day rebound and Waymo scaling optionality support a modest relative-strength call, but negative 30-day momentum keeps conviction low.
INTCUnderperform SPY55%A positive 5-day move is insufficient to offset a negative 30-day trend in a broadly weak semiconductor environment.
TSLAOutperform SPY55%Positive 30-day momentum supports a narrow relative-strength call, though negative 5-day momentum limits conviction.
MSFTUnderperform SPY54%Near-flat 5-day performance and a -1.14% 30-day return suggest modestly weaker relative performance than SPY absent clearer catalyst confirmation.
XOMOutperform SPY53%Positive 30-day energy momentum provides a slight relative edge versus SPY despite a negative recent week and loss versus cost.
^VIXUnderperform SPY53%With volatility already elevated over the past month and markets attempting to stabilize, VIX is modestly more likely to lag SPY over the horizon.
SPYOutperform SPY50%As the comparison benchmark, SPY is assigned a neutral reference call with no active relative-security preference.

The debate

Bull, bear and risk cases: bull 0.76 · bear 0.86 · risk 0.98
bull case · conviction 0.76

Buy or hold selectively: the evidence supports concentrating on leaders with clear relative strength rather than abandoning equities after a broad-market pullback. The S&P 500 is down 1.82% over 30 days, while several portfolio holdings are materially positive, led by META, AAPL, JNJ, PG, and V. AAPL is the strongest combination of momentum and identifiable catalysts, while META and JNJ provide additional leadership. The case is not universal across the portfolio. AMZN, HD, NVDA, UNH, and others are weaker, but the breadth of outperformers and the market's reported technology rebound argue for maintaining exposure and using weakness selectively rather than selling indiscriminately.

  • AAPL has strong relative strength: it is up 8.89% over 30 days and 1.79% over five days, versus the S&P 500 down 1.82% over 30 days and SPY down 3.06% over 30 days.
  • AAPL has multiple stated catalysts: CNBC reported an unusual long-term trading development, and Investor's Business Daily cited carrier promotions supporting iPhone 18 Pro sales. A current analyst price-target discussion also provides a potential sentiment catalyst, although no target level is supplied here.
  • META is the portfolio's strongest momentum leader, up 13.17% over 30 days and 4.49% over five days, demonstrating that high-quality growth exposure can still outperform despite market volatility.
  • JNJ is up 2.49% over 30 days and 0.35% over five days while the benchmark is negative; news identifies it as a trending stock and reports that it hit a 52-week high, supporting continued relative-strength leadership.
  • V and PG are also showing defensive relative strength: V is up 1.50% over 30 days and PG is up 1.91%, both outperforming the negative benchmark over the same period.
  • GOOGL has improving near-term momentum, rising 3.09% over five days after a nearly flat 30-day result. The constructive company evidence includes Waymo expanding to more domestic cities, entering international markets, and scaling operations, offering a longer-term optionality catalyst.
  • Market conditions are volatile rather than uniformly bearish: the VIX is up 21.05% over 30 days, but market reporting also describes technology-led recovery sessions as oil prices and Treasury yields eased. That favors owning relative-strength names while sizing risk carefully.
  • The evidence argues for selectivity, not blanket buying: AMZN is down 7.23% over 30 days, HD is down 10.83%, and NVDA is down 4.95%. These weaker names should not dilute conviction in the stronger leaders without a demonstrated momentum reversal.
bear case · conviction 0.86

The bearish case is that portfolio breadth and momentum are deteriorating while several holdings are already under water. Five holdings. HD, UNH, GOOGL, WMT, and XOM, are in the red, and multiple large positions are showing negative 30-day momentum. The weakness is not confined to one industry: retail, healthcare, energy, communications, technology, financials, and semiconductors all appear in the negative signals. Candidates look even more vulnerable, with severe 30-day declines in APLD, CRWV, AVGO, ASML, and CORZ. Given the S&P 500 is also down 1.8% over 30 days, the portfolio may be exposed to a broader risk-off trend rather than isolated company-specific weakness.

  • HD is down 7.3% from cost and has negative 30-day momentum of 10.8%, making it one of the clearest deterioration signals among current holdings.
  • UNH is down 6.6% and has negative 30-day momentum of 5.4%; the combination of realized weakness and continued fading momentum argues against assuming a near-term recovery.
  • AMZN has negative 30-day momentum of 7.2%, despite being only modestly above its stated average cost, raising the risk that the position's small gain erodes if weakness persists.
  • GOOGL and WMT are both below cost, while their 30-day momentum is also negative, 0.9% and 6.9%, respectively, indicating that these are not merely stale mark-to-market losses.
  • XOM is down 3.5% from cost, adding cyclical and sector exposure to the group of losing holdings; the evidence provides no positive momentum offset for this position.
  • Technology and semiconductor exposure is not immune: NVDA is down 4.95% over 30 days, TSM is down 2.71%, and MSFT is down 1.14%. The portfolio therefore remains exposed to weakening leadership even where prices are above average cost.
  • JPM is down 3.91% over 30 days, showing that financial exposure is also losing momentum; this broadens the downside case beyond consumer and technology holdings.
  • The candidate list contains substantially sharper deterioration than most current holdings: APLD is down 20.3%, CRWV 21.6%, AVGO 18.7%, ASML 13.3%, and CORZ 14.7% over 30 days. These are poor entry candidates while momentum is decisively negative.
  • GLW has negative 30-day momentum of 8.9%. The fund's cautionary memory specifically identifies risks around customer demand, customer financing and receivables, loss of significant customers, technology change, component performance, capital spending, margins, regulation, litigation, and acquisitions, creating multiple ways for weakness to continue.
  • The benchmark's 30-day return is negative 1.8%, so the market backdrop is already unfavorable. Holdings and candidates with materially worse momentum than the benchmark could carry disproportionate downside if the broad decline continues.
risk case · conviction 0.98

Capital risk is driven primarily by concentration: 64.5% of invested assets are in AAPL, JNJ, and AMZN, while 44.9% of invested assets are in Information Technology and 22.2% in Consumer Discretionary. Cash is a substantial 24.9% of the total portfolio, which reduces immediate market exposure but leaves the invested portion highly concentrated.

  • AAPL is the largest position at 28.9% of invested assets, making it the clearest single-name capital risk.
  • JNJ and AMZN add 18.6% and 17.0%, respectively; together with AAPL, the top three positions represent 64.5% of invested assets.
  • The portfolio has a pronounced sector tilt: Information Technology is 44.9%, Consumer Discretionary is 22.2%, and Health Care is 18.8% of invested assets.
  • AAPL, MSFT, NVDA, and TSM collectively account for 42.2% of invested assets, reinforcing the technology concentration beyond the largest position.
  • Cash at 24.9% provides a meaningful liquidity buffer and limits total-portfolio market exposure, but the remaining invested capital is concentrated rather than broadly distributed.
  • The provided benchmark has declined 1.82% over 30 days, but no comparable portfolio return is provided; therefore, relative performance or directional conclusions cannot be established from this evidence.

Cash thesis

Cash is 24.9% of portfolio value, just below the 25% threshold but still intentionally high. It is justified by concentrated exposure in AAPL, JNJ and AMZN, negative benchmark momentum, elevated volatility versus a month ago, and the lack of broad confirmation among technology and cyclical shares. Cash preserves flexibility to add only after trend confirmation rather than forcing purchases into falling candidates.

Full commentary

Bear case response

Accept: benchmark, SPY and QQQ momentum are negative, so broad risk appetite is not yet constructive. Accept: HD, UNH, WMT, XOM and GOOGL are below cost or have weak momentum; HD is the clearest unresolved deterioration and is trimmed. Accept: AMZN, NVDA, TSM, MSFT and JPM show soft 30-day momentum, so no additional purchases are justified. Accept: severely negative candidate momentum in APLD, CRWV, AVGO, ASML and CORZ makes them unsuitable for new entries. Accept: AAPL concentration is a material risk even though its relative strength is strong; this supports holding rather than adding. Reject a blanket liquidation approach: AAPL, META, JNJ, V and PG have demonstrated relative strength versus SPY, while the 24.9% cash balance allows defense without abandoning proven leaders. The response is selective risk reduction, not indiscriminate selling.

Market summary

The market remains uneven following Fed-related volatility: SPY is down 3.06% and QQQ down 3.74% over 30 days, while the VIX is up 21.05%. A reported intraday rebound in technology and easing yields improves near-term tone, but it does not yet reverse weak broader momentum. Leadership is narrow, with AAPL, META, JNJ, V and PG materially stronger than the benchmark, while consumer cyclicals, several semiconductors and parts of cloud/AI remain weak.

Portfolio assessment

The portfolio has meaningful gains in AAPL, JNJ, MSFT, META, V and MA, but is still concentrated: AAPL, JNJ and AMZN account for most invested capital, and AAPL alone is approximately 21.7% of total portfolio value. AMZN, HD, NVDA, TSM and GOOGL have weak or mixed 30-day momentum. A small trim of HD is warranted because its decline is persistent and unresolved; otherwise, core leaders should be held and no capital should be deployed into weak watchlist names.

Risk assessment

The principal risk is concentration rather than liquidity. AAPL, JNJ and AMZN dominate invested assets, while technology, AI-linked infrastructure and consumer-discretionary exposures could decline together if rates, consumer demand or AI spending expectations worsen. HD's -10.83% 30-day return, AMZN's -7.23%, GLW's -8.90%, and continued semiconductor softness demonstrate that weakness is broader than a single name. The cash reserve moderates total-portfolio drawdown risk but does not remove the impact of concentrated core positions.

← September 16, 2026 September 18, 2026 →