Glasshouse Fund
Decision journal

AI fund decision · August 18, 2026

Hold all positions and retain the 25.9% cash reserve. The decision reflects concentration risk, unavailable benchmark and equity momentum data, and a yield-driven market backdrop, while avoiding reactive sales in modestly weak holdings. Sources informing the restraint include risk_lesson:run_20260803T191113Z_b5f11a59:report_risk, risk_lesson:run_20260729T184341Z_54100baf:report_risk, trade:run_20260812T183945Z_0d1bf062:AAPL:SELL, and trade:run_20260807T183247Z_07bc8571:AAPL:BUY.

Outlook: NEUTRALPortfolio $1,035,159Cash 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 (16 outperform, 18 underperform); 0 became trades. These are the fund's calibration record.

beat SPYlag SPYAAPL: underperform at 56%AMD: underperform at 59%AMZN: outperform at 56%APLD: underperform at 58%ASML: outperform at 55%AVGO: outperform at 56%BE: underperform at 57%CAT: outperform at 54%CORZ: underperform at 58%CRWV: underperform at 57%GLW: outperform at 54%GOOGL: underperform at 57%HD: underperform at 55%INTC: underperform at 58%IREN: underperform at 58%JNJ: outperform at 58%JPM: outperform at 59%MA: outperform at 57%META: underperform at 56%MSFT: outperform at 56%MU: underperform at 58%NVDA: underperform at 56%ORCL: outperform at 54%PG: outperform at 57%QQQ: underperform at 58%SNDK: underperform at 58%SPY: outperform at 50%TSLA: underperform at 59%TSM: underperform at 56%UNH: outperform at 54%V: outperform at 57%WMT: outperform at 57%XOM: outperform at 59%^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
AMDUnderperform SPY59%Chip stocks are under pressure amid bond-market anxiety, with no verified momentum data to support relative strength.
JPMOutperform SPY59%Higher yields can support bank economics, and the shares are materially above recorded cost, though macro stress remains a risk.
TSLAUnderperform SPY59%High-duration consumer-discretionary exposure is vulnerable to rising yields, with no supplied catalyst supporting relative strength.
XOMOutperform SPY59%Rising oil referenced in market headlines provides a relative tailwind versus the broader market.
APLDUnderperform SPY58%Recent commentary references the stock's worst month in a year, and no confirming fundamental evidence offsets elevated AI-infrastructure risk.
CORZUnderperform SPY58%The name has high sensitivity to speculative AI and digital-infrastructure financing conditions as yields rise.
INTCUnderperform SPY58%Reported sector weakness in Intel and AMD, combined with chip pressure from rising yields, is unfavorable versus SPY.
IRENUnderperform SPY58%Risk-sensitive infrastructure exposure lacks supplied fundamental confirmation during a higher-yield market stress episode.
JNJOutperform SPY58%Defensive healthcare characteristics and MONARCH clearance support relative resilience, despite valuation concerns.
MUUnderperform SPY58%Semiconductor shares are specifically pressured by bond anxiety, and no supplied company-specific evidence offsets that risk.
QQQUnderperform SPY58%The growth-heavy index is especially exposed to rising yields and reported semiconductor weakness.
SNDKUnderperform SPY58%The semiconductor backdrop is unfavorable under global bond-market anxiety, with no supplied offsetting evidence.
BEUnderperform SPY57%Supply-chain allegations and a highly volatile operating backdrop create greater uncertainty than the broad market.
CRWVUnderperform SPY57%Limited supplied company-specific evidence and risk-sensitive AI infrastructure exposure favor caution versus SPY.
GOOGLUnderperform SPY57%The position is below cost and Waymo expansion is longer-duration optionality rather than near-term proof of relative outperformance.
MAOutperform SPY57%A strong gain versus cost and durable payments economics support relative performance, with consumer spending as the key risk.
PGOutperform SPY57%Defensive consumer-staples exposure should be comparatively resilient in a higher-yield, risk-off equity session.
VOutperform SPY57%Payments quality and the position's gain versus cost support a modest relative-performance case, subject to consumer-spending risk.
WMTOutperform SPY57%Defensive retail characteristics are comparatively favorable if yields and macro uncertainty continue to pressure cyclicals.
AAPLUnderperform SPY56%AI supply-chain headwinds and exceptional portfolio concentration outweigh the speculative foldable-device catalyst versus SPY.
AMZNOutperform SPY56%AWS-driven long-term optimism supports relative resilience, though higher yields and consumer-discretionary exposure limit conviction.
AVGOOutperform SPY56%AI exposure remains supportive despite recent VMware concerns and broader semiconductor volatility.
METAUnderperform SPY56%The shares are below recorded cost and no supplied catalyst establishes a near-term relative-performance advantage.
MSFTOutperform SPY56%Cloud and software quality support resilience, but China-related Windows risk and higher yields reduce conviction.
NVDAUnderperform SPY56%AI demand remains constructive, but chip-sector pressure, valuation sensitivity, and prior risk memory favor caution versus SPY.
TSMUnderperform SPY56%Structural AI demand is favorable, but current sector-wide chip pressure and valuation sensitivity dominate near term.
ASMLOutperform SPY55%AI semiconductor-capex leadership is constructive, but the premium valuation and yield-sensitive market keep confidence low.
HDUnderperform SPY55%Reaffirmed guidance is constructive, but management's frozen-housing-market description leaves cyclical demand risk elevated.
^VIXUnderperform SPY55%The VIX has declined over five and 30 days; relative to SPY this implies a lower probability of sustained volatility outperformance, though yields could reverse the trend.
CATOutperform SPY54%Infrastructure and AI-related demand themes are constructive, though the cyclical valuation backdrop limits conviction.
GLWOutperform SPY54%Technology-infrastructure demand provides a possible tailwind, although no current return data validates a stronger call.
ORCLOutperform SPY54%Enterprise-cloud and AI infrastructure positioning may support relative demand, but evidence is insufficient for high conviction.
UNHOutperform SPY54%Healthcare defensiveness and expanded member coaching access may support resilience, despite the shares being modestly below cost.
SPYOutperform SPY50%As the benchmark, SPY is assigned neutral mechanical relative performance with no directional edge.

The debate

Bull, bear and risk cases: bull 0.62 · bear 0.62 · risk 0.98
bull case · conviction 0.62

The strongest bullish case is to hold the portfolio and selectively add to its proven winners rather than react to a macro-driven pullback. Several holdings are materially above cost, including MA, MSFT, JPM, V, JNJ, AMZN, NVDA, and HD, demonstrating positive portfolio-level relative strength even as yields, oil, and bond-market stress pressure equities. The volatility backdrop is also improving: VIX is down 6.7% over five days and 13.6% over 30 days. Near-term catalysts include Home Depot reaffirming fiscal-2026 guidance, J&J's MONARCH clearance, Apple's potential foldable-iPhone opportunity, and Alphabet's Waymo expansion and commercialization. The evidence is not strong enough for an aggressive portfolio-wide purchase because all reported equity momentum and benchmark returns are unavailable, but it supports maintaining exposure and emphasizing the strongest holdings.

  • Cross-sectional strength is favorable: MA, MSFT, JPM, V, JNJ, AMZN, NVDA, and HD are all currently above their recorded average costs, with MA, MSFT, JPM, V, and JNJ among the clearest winners.
  • Risk sentiment has improved on the supplied volatility measure: VIX declined 6.7% over five days and 13.6% over 30 days, a constructive backdrop despite headlines about rising yields and oil.
  • HD reaffirmed fiscal-2026 guidance despite describing housing conditions as frozen, reducing the risk of a near-term guidance shock.
  • JNJ has a tangible product catalyst: MONARCH clearance is described as lifting momentum, although valuation concerns remain.
  • Alphabet has longer-duration upside optionality through Waymo, which is expanding domestically, entering international markets, and scaling operations.
  • Apple has a potential product catalyst in a foldable iPhone, while its large installed ecosystem remains a reason to hold; however, the evidence also flags AI supply-chain headwinds.
  • Formal five-day and 30-day price momentum for the stocks and the S&P 500 benchmark are unavailable, so claims of market-beating momentum cannot be verified from this dataset.
  • The appropriate bullish stance is selective accumulation or holding, not indiscriminate buying: UNH, GOOGL, and META are below recorded average cost, and MSFT faces a reported China-related Windows risk.
bear case · conviction 0.62

The clearest bearish case is that weakness is already appearing in UNH, GOOGL, and META, while the portfolio lacks a usable benchmark comparison to confirm that the broader holdings are outperforming. The available momentum evidence is also cautionary: the VIX is down 13.6% over 30 days, and the signal is explicitly characterized as fading momentum. With no positive benchmark data or candidate-specific evidence provided, there is insufficient support for adding risk, particularly to names already showing losses.

  • UNH is below its average cost by approximately 1.9%, indicating current weakness rather than immediate follow-through from the holding.
  • GOOGL is below its average cost by approximately 2.7%, making it the weakest of the identified red positions and a direct source of portfolio downside.
  • META is below its average cost by approximately 3.0%, the largest reported loss among the holdings with downside signals.
  • The 30-day VIX change is -13.6% and is labeled as fading momentum; this provides a caution signal rather than evidence of strengthening risk conditions.
  • The S&P 500 benchmark return and price fields are unavailable, so there is no reliable evidence that the portfolio or its winners are outperforming the market.
  • The evidence does not establish positive momentum for the many holdings currently above cost; gains could therefore be vulnerable if the limited breadth of weakness expands.
  • No candidates are specified, and no candidate-specific evidence supports initiating additional positions.
risk case · conviction 0.98

The largest capital risks are concentration in three individual holdings and a substantial tilt toward Information Technology and Consumer Discretionary. AAPL, AMZN, and JNJ represent 71.4% of invested assets, while the two largest sectors represent 71.0% of invested assets. Cash at 25.9% provides a material buffer, but the invested portfolio remains highly dependent on a small number of positions and sectors.

  • AAPL is 27.3% of invested assets, the largest single-position exposure.
  • AMZN is 24.9% and JNJ is 19.2%; together with AAPL, these three holdings comprise 71.4% of invested assets.
  • Information Technology is 40.2% and Consumer Discretionary is 30.8% of invested assets, creating significant sector concentration.
  • AAPL, MSFT, NVDA, and GOOGL collectively account for 47.1% of invested assets, reinforcing concentration in technology-related exposures.
  • The remaining holdings are comparatively small, limiting their ability to offset losses in the largest positions.
  • Cash is 25.9% of the total portfolio and reduces exposure to invested-asset losses, but it does not diversify the invested holdings themselves.
  • The benchmark return and current level are unavailable, so no evidence-based relative-risk or tracking assessment can be made.

Cash thesis

Cash is 25.9% of portfolio value, above the 25% threshold. This reserve is intentional: it offsets substantial single-name and technology/consumer-discretionary concentration, preserves flexibility during a yield-driven market pullback, and avoids deploying capital without usable stock momentum, benchmark-relative returns, or sufficiently specific valuation evidence. The opportunity cost is acceptable until market breadth or company-specific confirmation improves.

Full commentary

Bear case response

Accept: GOOGL, META, and UNH are below recorded cost, which confirms modest current weakness and argues against adding to them today. Accept: unavailable S&P 500 and individual-stock return data prevent a reliable claim of benchmark-relative outperformance, so above-cost positions should not be treated as confirmed momentum leaders. Accept in part: the lower VIX is ambiguous; it may reflect easing fear, but it does not override the yield and oil pressure or establish durable equity breadth. Reject the implication that the red positions require immediate selling: losses are limited, no thesis-breaking company-specific development is established, and selling solely on small drawdowns would conflict with the mandate to avoid overtrading. Accept: no new candidate has adequate supplied valuation, earnings, or momentum evidence for initiation. Accordingly, the response is to hold existing positions and preserve cash rather than add risk or execute reactive sales.

Market summary

The immediate backdrop is cautious: equity futures and chip shares are under pressure as global bond selling, higher yields, and oil weigh on risk assets. The VIX is lower over both five and 30 days, but equity and benchmark return series are unavailable, so breadth and relative-strength confirmation cannot be established. This supports retaining quality core exposure while avoiding fresh risk-taking.

Portfolio assessment

The portfolio is up modestly in aggregate and retains profitable core positions in JNJ, AAPL, MSFT, NVDA, AMZN, HD, JPM, V, and MA. However, AAPL ($209.4k), AMZN ($190.9k), and JNJ ($147.0k) remain dominant exposures, while GOOGL, META, and UNH are below cost. No trade is warranted today: prior recent-trade memory shows repeated AAPL adjustments, and current evidence does not justify reversing or adding to that exposure.

Risk assessment

Concentration is the principal risk. AAPL, AMZN, and JNJ account for approximately 71.4% of invested assets, and technology plus consumer-discretionary exposure is approximately 71.0% of invested assets. Higher yields may pressure long-duration technology valuations, while frozen housing conditions remain a cyclical risk for HD. The small positions in financials, payments, healthcare, and other names do not materially offset losses if the largest holdings weaken together. Recent risk lessons favor avoiding overtrading and using incremental trims only when deterioration is clearly evidenced.

← August 17, 2026