Glasshouse Fund
Decision journal

AI fund decision · August 17, 2026

Hold existing core positions and retain the 26.1% cash reserve. Do not add to concentrated AAPL or AMZN, do not average down in UNH, GOOGL, META, or other laggards, and do not chase high-volatility AI-infrastructure names after sharp rallies. This preserves participation in MSFT, NVDA, AMZN, and payments strength while respecting concentration and prior risk-management lessons.

Outlook: NEUTRALPortfolio $1,025,785Cash 26.1%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 (15 outperform, 19 underperform); 0 became trades. These are the fund's calibration record.

beat SPYlag SPYAAPL: underperform at 61%AMD: underperform at 57%AMZN: outperform at 59%APLD: outperform at 56%ASML: outperform at 64%AVGO: outperform at 55%BE: underperform at 54%CAT: underperform at 66%CORZ: underperform at 63%CRWV: outperform at 55%GLW: underperform at 68%GOOGL: underperform at 58%HD: underperform at 56%INTC: underperform at 53%IREN: outperform at 57%JNJ: underperform at 54%JPM: outperform at 63%MA: outperform at 62%META: underperform at 65%MSFT: outperform at 67%MU: outperform at 60%NVDA: outperform at 65%ORCL: outperform at 61%PG: underperform at 56%QQQ: underperform at 55%SNDK: underperform at 53%SPY: outperform at 50%TSLA: underperform at 69%TSM: underperform at 54%UNH: underperform at 67%V: outperform at 58%WMT: underperform at 52%XOM: outperform at 57%^VIX: underperform at 64%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
TSLAUnderperform SPY69%The 13.3% 30-day decline reflects substantial relative weakness and company-specific uncertainty despite a positive 5-day move.
GLWUnderperform SPY68%Negative 30-day momentum and disclosed execution, customer, financing, and margin risks create an unfavorable risk-adjusted setup versus SPY.
MSFTOutperform SPY67%Its 26.7% 30-day gain far exceeds SPY and supports AI/cloud leadership, although the recent pullback and chip-gap concern temper conviction.
UNHUnderperform SPY67%Negative 5-day and 30-day momentum, combined with a price below portfolio cost, make this the clearest current holding weakness.
CATUnderperform SPY66%The 7.9% 30-day decline reflects established relative weakness despite a modest recent bounce.
METAUnderperform SPY65%The 10.2% 30-day decline and position below cost indicate pronounced relative weakness despite limited 5-day stabilization.
NVDAOutperform SPY65%Positive 5-day and 30-day returns plus OpenAI ecosystem validation support continued AI-demand leadership.
ASMLOutperform SPY64%Strong positive 5-day and 30-day returns indicate durable semiconductor-equipment relative strength.
^VIXUnderperform SPY64%The 5-day and 30-day declines in volatility are consistent with a risk-on regime in which VIX should lag SPY.
CORZUnderperform SPY63%Positive short-term momentum is insufficient to overcome an 11.1% 30-day decline and elevated speculative volatility.
JPMOutperform SPY63%Positive 5-day momentum and an 8.5% 30-day gain above SPY support continued financial-sector relative strength.
MAOutperform SPY62%Positive short-term momentum and a 5.9% 30-day gain modestly above SPY support continued payments leadership.
AAPLUnderperform SPY61%Negative 5-day and 30-day momentum indicates weaker relative performance than SPY until product and AI catalysts translate into renewed price leadership.
ORCLOutperform SPY61%A 14.4% 30-day gain materially above SPY supports cloud and enterprise-software relative strength despite flat recent trading.
MUOutperform SPY60%Strong 5-day momentum and favorable memory-industry headlines support relative strength, though the move is extended.
AMZNOutperform SPY59%AWS and AI-demand commentary support the 30-day gain above SPY, though the recent 5-day pullback lowers conviction.
GOOGLUnderperform SPY58%Berkshire-related interest and Waymo optionality are constructive, but negative 5-day and 30-day momentum remain weaker than SPY.
VOutperform SPY58%Consistent positive 5-day and 30-day returns support modest payments-sector outperformance versus SPY.
AMDUnderperform SPY57%The 30-day decline despite a strong 5-day rebound suggests volatile recovery trading rather than confirmed relative leadership.
IRENOutperform SPY57%Strong 5-day and 30-day gains tied to AI-cloud contract momentum support relative strength, albeit with elevated volatility.
XOMOutperform SPY57%Strong 30-day energy momentum and current oil-related geopolitical tension support relative performance, though recent momentum is flat.
APLDOutperform SPY56%Positive 5-day and 30-day momentum supports continued AI-infrastructure strength, but valuation concerns make this a low-confidence call.
HDUnderperform SPY56%A materially negative 5-day return, subdued 30-day gain, and CEO medical-leave uncertainty imply limited near-term relative strength.
PGUnderperform SPY56%Negative 5-day and 30-day returns show weaker momentum than SPY in the current risk-on market.
AVGOOutperform SPY55%The positive 30-day return supports the AI infrastructure thesis, although the negative 5-day return signals near-term consolidation.
CRWVOutperform SPY55%Exceptional 5-day and 30-day momentum supports relative strength, but insider-sale news and high volatility constrain confidence.
QQQUnderperform SPY55%The 2.7% 30-day gain trails SPY's 3.6% return, indicating broad technology has lagged despite strong individual AI leaders.
BEUnderperform SPY54%A sharp 5-day rebound does not yet offset negative 30-day momentum and the name remains higher-risk than SPY.
JNJUnderperform SPY54%Defensive quality and positive 30-day performance support stability, but modest momentum likely trails a constructive SPY environment.
TSMUnderperform SPY54%Positive but modest 5-day and 30-day gains lag SPY, leaving insufficient evidence of near-term outperformance.
INTCUnderperform SPY53%The 5-day rebound is constructive, but flat-to-negative 30-day momentum does not yet establish a durable edge over SPY.
SNDKUnderperform SPY53%The sharp 5-day surge is momentum-positive, but negative 30-day performance and high cyclicality make sustained outperformance uncertain.
WMTUnderperform SPY52%Positive momentum is constructive, but its modest 30-day gain trails SPY and limits expected relative upside.
SPYOutperform SPY50%As the comparison benchmark, SPY is assigned neutral benchmark-level expected performance.

The debate

Bull, bear and risk cases: bull 0.76 · bear 0.87 · risk 0.98
bull case · conviction 0.76

Buy or hold selectively: the strongest holdings are demonstrating clear relative strength against a constructive market backdrop, led by Microsoft, Nvidia, Amazon, JPMorgan, and Mastercard. The S&P 500 gained 4.6% over 30 days while volatility declined sharply, with the VIX down 17.0% over the same period. Several portfolio holdings materially outperformed that benchmark, and the AI/cloud ecosystem continues to provide the most compelling catalyst set. The evidence supports adding to proven relative-strength leaders rather than broadly buying laggards.

  • MSFT is the clearest momentum leader in the portfolio: its 30-day gain was 26.7%, far above the S&P 500's 4.6% return and QQQ's 2.7%. The cloud business remains a central investor focus, despite near-term concern about a reported 2.2 million chip gap.
  • NVDA gained 10.6% over 30 days and 3.5% over five days, demonstrating both intermediate and recent strength. Its reported deal with OpenAI, under which OpenAI becomes a Nvidia tenant, is a potentially important catalyst for continued AI demand and ecosystem validation.
  • AMZN rose 6.2% over 30 days, outperforming the broad market despite a weak recent five-day period. That combination supports holding through short-term volatility while maintaining exposure to a major market leader.
  • JPMorgan gained 8.5% over 30 days and Mastercard gained 5.9%, both exceeding the S&P 500. Their positive five-day momentum, 0.8% for JPM and 1.1% for MA, indicates that the relative strength is not entirely stale.
  • Visa also remains constructive, with positive five-day and 30-day momentum of 1.0% and 2.0%, respectively. The payments holdings provide diversification beyond the AI trade while still participating in market strength.
  • The market backdrop is supportive for maintaining risk exposure: the S&P 500 rose 4.6% over 30 days, SPY rose 3.6%, QQQ rose 2.7%, and the VIX declined 17.0%. This suggests improved risk appetite even as traders weigh AI optimism against geopolitical tensions.
  • GOOGL has a specific positive catalyst: Berkshire Hathaway reportedly increased its Alphabet stake by 83%, with Buffett saying the investment was his idea. Alphabet's disclosed Waymo expansion and international scaling also provide a longer-term optionality catalyst, although its own 30-day momentum remains weak at -1.9%.
  • The evidence argues for selectivity rather than indiscriminate buying. AAPL, GOOGL, META, UNH, and HD have negative or subdued 30-day momentum, while MSFT, NVDA, AMZN, JPM, and MA show the strongest combination of momentum and relative strength. Apple's China custom-AI work is a constructive development, but its 30-day return of -3.5% does not yet confirm market leadership.
bear case · conviction 0.87

The clearest bearish case is concentrated in holdings and candidates showing both absolute weakness and deteriorating momentum while the S&P 500 gained 4.6% over the last 30 days. UNH, GOOGL, and META are already below cost, and all three also have negative 30-day momentum; AAPL is near cost but has declined 3.5% over 30 days. This creates evidence of weakening relative performance rather than broad market weakness. Several candidates are materially worse, including TSLA, CORZ, CAT, and GLW, while GLW also carries explicit risks around customer losses, financing, technology change, margins, capital spending, litigation, regulation, and product performance.

  • UNH is down 2.8% from average cost and has negative 30-day momentum of 6.4%, making it the strongest current holding warning.
  • GOOGL is down 2.3% from average cost and has negative 30-day momentum of 1.9%; META is effectively below cost and has sharply negative 30-day momentum of 10.2%.
  • AAPL remains above average cost but its 30-day momentum is negative 3.5%, suggesting the position's recent gains are losing support.
  • The portfolio's identified weak holdings contrast with a 4.6% 30-day gain in the S&P 500, indicating that these signals cannot be dismissed solely as market-wide weakness.
  • Candidate risk is more severe: TSLA is down 13.3% over 30 days, CORZ 11.1%, GLW 9.3%, and CAT 7.9%; entering these names would amount to buying into pronounced deterioration.
  • GLW's disclosed risk set includes loss of significant customers, inability to pace capital spending with demand, margin execution failure, technology change, product issues, litigation, regulation, and customer financing or payment problems, risks that compound its 9.3% recent decline.
  • Negative momentum also appears across AMD, BE, INTC, PG, and SNDK, implying that the weakness is not isolated to one industry or one candidate.
risk case · conviction 0.98

The primary capital risk is concentration: 52.2% of invested assets is in AAPL and AMZN, while 71.0% is concentrated in the top three positions once JNJ is included. This is reinforced by a substantial tilt toward Information Technology and Consumer Discretionary, which together represent 71.2% of reported exposure. Cash at 26.1% provides a meaningful liquidity and drawdown buffer, but it does not diversify the invested holdings; the deployed capital remains heavily dependent on a small number of positions and sectors.

  • AAPL is the largest position at 27.1% of invested assets, creating the single biggest position-specific capital exposure.
  • AMZN is 25.1%; combined with AAPL, the two positions account for 52.2% of invested assets.
  • JNJ is 18.8%, bringing the top three positions to 71.0% of invested assets.
  • Information Technology is 40.2% and Consumer Discretionary is 31.0%, producing a 71.2% combined sector concentration.
  • Health Care adds another 19.0%, so the three largest reported sectors account for 90.2% of exposure.
  • The remaining sectors are small: Communication Services is 7.7% and Financials are 2.1%, limiting diversification outside the dominant sectors.
  • Cash is 26.1% of the portfolio and reduces overall invested exposure, but it does not offset the concentration of the capital that is invested.
  • The benchmark's 30-day return is provided, but no portfolio return is provided, so relative performance or benchmark-relative risk cannot be assessed from this evidence.

Cash thesis

Cash is 26.1% of portfolio value and is intentionally retained. It cushions drawdown risk from concentrated AAPL, AMZN, JNJ, and technology/consumer-discretionary exposure; preserves flexibility if geopolitical risks create a broader pullback; and avoids deploying capital into lagging names or chasing sharp AI-infrastructure rallies. The cash balance is an active risk-control allocation rather than an absence of conviction.

Full commentary

Bear case response

Accept: UNH is the clearest holding warning, with a 2.8% loss versus cost and a 6.4% 30-day decline; its very small position should not be increased without evidence of reversal. Accept: GOOGL and META are below cost with negative 30-day momentum, and AAPL's 3.5% 30-day decline is weak relative to a rising benchmark; none should receive new capital today. Accept: negative returns in TSLA, CORZ, CAT, GLW, and several other watchlist names argue against initiating positions, with GLW's disclosed customer, margin, capital-spending, technology, legal, and regulatory risks adding to the case for avoidance. Partly reject the implication that all risk exposure should be reduced: MSFT, NVDA, AMZN, JPM, MA, and V retain comparatively better momentum or business support, so wholesale selling would sacrifice demonstrated leaders. The appropriate response is selective defense, hold stronger existing positions, avoid laggards and speculative momentum chases, and maintain cash, rather than indiscriminate selling.

Market summary

The broad backdrop remains constructive: SPY gained 3.6% over 30 days, the S&P 500 benchmark gained 4.6%, and the VIX declined 17.0%. However, the current session is softer amid Iran-related geopolitical tension and higher oil, while leadership remains narrow and concentrated in AI/cloud, semiconductors, and selected financials. This supports retaining quality exposure but does not justify chasing extended winners or averaging down in lagging holdings.

Portfolio assessment

The portfolio is liquid and profitable in several core positions, particularly MSFT, NVDA, AMZN, JNJ, and payment networks. However, AAPL, AMZN, and JNJ remain very large positions, while AAPL, GOOGL, META, and UNH have weaker recent relative momentum. No transaction is warranted today: the August 12 AAPL trim already addressed some concentration, and the current evidence does not justify reversing that discipline by adding to AAPL or AMZN after mixed near-term action.

Risk assessment

The principal risk is concentration rather than market-wide volatility. AAPL and AMZN represent 52.2% of invested assets, and AAPL, AMZN, and JNJ together represent 71.0% of invested assets. Technology and consumer discretionary exposure are also dominant. A common reversal in AI spending, cloud expectations, consumer demand, or large-cap technology valuation could affect multiple major positions simultaneously. The portfolio should continue to avoid enlarging its largest positions until momentum and diversification improve.

← August 14, 2026 August 18, 2026 →