Glasshouse Fund
Decision journal

AI fund decision · August 20, 2026

Maintain a neutral, low-turnover stance. Hold core relative-strength positions, do not add to laggards or chase extended winners, and preserve the elevated cash reserve until concentration is reduced through deliberate future rebalancing or clearer entry signals emerge.

Outlook: NEUTRALPortfolio $1,037,281Cash 25.8%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 (12 outperform, 22 underperform); 0 became trades. These are the fund's calibration record.

beat SPYlag SPYAAPL: underperform at 59%AMD: underperform at 67%AMZN: outperform at 62%APLD: underperform at 61%ASML: underperform at 58%AVGO: underperform at 69%BE: underperform at 70%CAT: underperform at 65%CORZ: underperform at 73%CRWV: underperform at 65%GLW: underperform at 59%GOOGL: underperform at 60%HD: underperform at 56%INTC: underperform at 68%IREN: underperform at 57%JNJ: outperform at 68%JPM: outperform at 59%MA: outperform at 60%META: underperform at 72%MSFT: outperform at 61%MU: outperform at 59%NVDA: outperform at 57%ORCL: outperform at 55%PG: underperform at 57%QQQ: underperform at 55%SNDK: outperform at 58%SPY: outperform at 50%TSLA: underperform at 67%TSM: underperform at 60%UNH: underperform at 68%V: outperform at 55%WMT: underperform at 55%XOM: outperform at 64%^VIX: underperform at 58%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
CORZUnderperform SPY73%CORZ's 10.9% 30-day loss and 5.4% five-day loss outweigh positive data-center-plan headlines.
METAUnderperform SPY72%META's 17.8% 30-day loss and 8.2% five-day decline are the strongest negative momentum signals among current holdings.
BEUnderperform SPY70%BE fell 12.5% over five days and has not demonstrated relative strength versus SPY.
AVGOUnderperform SPY69%AVGO's 13.2% five-day decline and negative 30-day performance indicate unfavorable momentum despite dip-buying commentary.
INTCUnderperform SPY68%INTC's 11.2% five-day decline and negative 30-day return indicate pronounced unfavorable momentum.
JNJOutperform SPY68%JNJ's 9.4% 30-day and 4.3% five-day gains provide strong defensive relative strength outside technology.
UNHUnderperform SPY68%UNH is below cost and down 8.2% over 30 days, with governance-related concerns adding to stock-specific risk.
AMDUnderperform SPY67%AMD is down 6.9% over 30 days and 3.4% over five days, indicating sustained relative weakness.
TSLAUnderperform SPY67%TSLA's 10.2% 30-day decline outweighs its recent five-day bounce and indicates unstable relative momentum.
CATUnderperform SPY65%CAT is down 4.5% over five days and 6.8% over 30 days, a clear weak trend against a positive market.
CRWVUnderperform SPY65%Despite a strong 30-day return, CRWV's 14.5% five-day drop and reported debt and dilution risk make it less attractive than SPY.
XOMOutperform SPY64%XOM's 13.6% 30-day and 4.6% five-day gains, alongside higher-oil-risk commentary, support strong relative momentum.
AMZNOutperform SPY62%AMZN's 6.4% 30-day gain exceeds SPY, supported by constructive AWS-growth commentary, although five-day momentum is muted.
APLDUnderperform SPY61%A 6.8% 30-day gain is offset by a sharp 7.8% five-day decline, signaling high volatility and weakening near-term relative momentum.
MSFTOutperform SPY61%MSFT's 20.7% 30-day advance and AI-efficiency narrative support leadership, tempered by extension and recent five-day weakness.
GOOGLUnderperform SPY60%Waymo commercialization is a long-term positive, but current negative 30-day momentum and a below-cost position point to near-term relative weakness.
MAOutperform SPY60%MA has outperformed over both 30 days and five days, supporting a modest relative-strength view.
TSMUnderperform SPY60%TSM is down 4.3% over five days and has only marginal 30-day gains, trailing the benchmark's stronger 30-day performance.
AAPLUnderperform SPY59%A short-term rebound is constructive, but negative 30-day momentum and concentration-sensitive technology valuation risk leave AAPL modestly less favorable than SPY.
GLWUnderperform SPY59%Negative five-day and 30-day momentum combine with disclosed demand, margin, customer and execution risks.
JPMOutperform SPY59%JPM's 4.1% 30-day gain exceeds SPY, although the recent five-day pullback moderates conviction.
MUOutperform SPY59%MU's 9.8% 30-day gain remains well ahead of SPY despite a modest recent decline.
ASMLUnderperform SPY58%ASML has negative five-day and 30-day returns, with no supplied catalyst sufficient to establish a reversal.
SNDKOutperform SPY58%SNDK's positive five-day and 11.2% 30-day returns indicate favorable relative momentum.
^VIXUnderperform SPY58%The VIX is down 11.0% over 30 days and is likely to lag equities if market conditions remain orderly, although yield shocks are a risk.
IRENUnderperform SPY57%Strong 30-day gains are offset by a weak five-day trend and elevated volatility associated with AI and crypto-infrastructure expectations.
NVDAOutperform SPY57%NVDA's 4.9% 30-day gain and reported AI-demand strength support modest outperformance, but earnings event risk limits confidence.
PGUnderperform SPY57%PG has negative 30-day momentum and limited near-term price confirmation relative to SPY.
HDUnderperform SPY56%HD is only modestly above cost, has slightly negative 30-day momentum, and faces a frozen housing-market backdrop.
ORCLOutperform SPY55%ORCL's strong 30-day return and AI-backlog narrative are favorable, though its 7.9% five-day decline leaves conviction low.
QQQUnderperform SPY55%QQQ has only modestly outperformed over 30 days and is declining over five days as rising yields pressure technology breadth.
VOutperform SPY55%V's price trend is nearly flat but positive, and payments exposure offers steadier relative characteristics than high-beta technology.
WMTUnderperform SPY55%WMT has slightly negative five-day and 30-day returns, providing insufficient evidence of near-term relative strength.
SPYOutperform SPY50%SPY is the benchmark reference; the forced directional classification is neutral at 0.50 confidence.

The debate

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

The strongest case is to hold and selectively add to the portfolio’s clear relative-strength leaders rather than make broad changes. The S&P 500 gained 2.6% over 30 days, while MSFT (+20.7%), JNJ (+9.4%), AMZN (+6.4%), NVDA (+4.9%), JPM (+4.1%), and MA (+4.0%) outperformed. This breadth spans technology, healthcare, consumer, semiconductors, banking, and payments, providing evidence of sustained leadership rather than reliance on a single theme. Near-term catalysts include Microsoft’s AI-chip efficiency gains, continued strong AI demand ahead of NVDA earnings next week, Apple’s reported foldable-iPhone and camera-equipped AirPods opportunities, and Alphabet’s expanding Waymo commercialization. The portfolio also retains approximately $268,000 in cash, allowing disciplined additions on weakness. The bullish case is strongest for MSFT, JNJ, AMZN, NVDA, and MA, while weaker recent performers such as GOOGL, META, UNH, and AAPL warrant more selective sizing rather than aggressive averaging. Risks are real: bond yields are rising, market commentary is warning of forward trouble, and several holdings have recently lagged. Conviction is therefore positive but not extreme.

  • MSFT is the clearest momentum leader: its 30-day return of 20.7% was roughly 18 percentage points above the S&P 500, supported by news of AI-chip efficiency gains and recent institutional position buying.
  • JNJ gained 9.4% over 30 days and 4.3% in the last five sessions, materially outperforming the benchmark while offering portfolio leadership outside technology.
  • AMZN rose 6.4% over 30 days, more than twice the S&P 500’s gain, giving the portfolio consumer-sector momentum despite only modest five-day movement.
  • NVDA is up 4.9% over 30 days versus the S&P 500’s 2.6%; the visible catalyst is next week’s earnings, with reported commentary describing AI demand as 'rock-solid.'
  • JPM (+4.1%) and MA (+4.0%) also beat the benchmark over 30 days, confirming relative strength in financials and payments; MA additionally gained 1.2% over five days.
  • AAPL’s five-day gain of 3.8% shows a possible short-term rebound despite a weak 30-day return of -4.8%; reported product possibilities for a foldable iPhone and camera-equipped AirPods provide optionality.
  • GOOGL’s reported 10-K evidence identifies Waymo as expanding domestically, entering international markets, and scaling operations, offering a commercialization catalyst even though the stock’s 30-day return is -2.7%.
  • The portfolio has substantial liquidity, about $268,000 in cash, so the committee can hold core winners and add selectively without being forced into concentrated buying.
  • The principal counterweight is macro risk: market news cites jumping bond yields and warnings that the bond market may signal trouble ahead; META, UNH, GOOGL, and AAPL have also lagged over 30 days.
bear case · conviction 0.84

The clearest bearish signal is deteriorating relative momentum: the S&P 500 gained 2.65% over the last 30 days while several holdings and many candidates declined. Weakness is concentrated in names already below cost and in high-profile growth or cyclical exposures, creating risk that recent losses deepen rather than mean-revert. The evidence does not justify adding to broad market winners where momentum data is unavailable, and it argues for avoiding the weakest names until the trend improves.

  • UNH is already 4.28% below cost and has the portfolio's sharpest documented holding-level momentum deterioration, down 8.21% over 30 days.
  • META is 4.51% below cost and down 17.83% over 30 days, the strongest negative momentum signal among the listed holdings.
  • GOOGL is 3.26% below cost and down 2.75% over 30 days, indicating both current loss and continued weakening.
  • AAPL is profitable relative to its average cost but has fallen 4.85% over 30 days, so the position is exposed to a reversal of recent gains.
  • HD is also showing negative 30-day momentum of 1.07%, while its current price is only modestly above average cost, leaving less demonstrated cushion against further weakness.
  • The candidate list is broadly unfavorable: CORZ is down 10.94%, TSLA 10.21%, AMD 6.89%, CAT 6.78%, AVGO 3.20%, GLW 3.74%, and ASML 1.73% over 30 days.
  • GLW combines negative momentum with explicitly disclosed risks around customer demand, margin improvement, loss of significant customers, product performance, litigation, regulation, and the ability to pace capital spending, multiple ways for weakness to persist.
  • The benchmark's 2.65% gain makes the deterioration in META, UNH, GOOGL, AAPL, and the leading candidates more concerning: this is underperformance during a rising-market period, not merely weakness explained by a broad market decline.
  • Several current holdings have no supplied momentum evidence, so their apparent gains versus average cost should not be treated as confirmation of ongoing strength; the available evidence is asymmetric toward caution.
risk case · conviction 0.98

The dominant capital risk is concentration: 71.6% of invested assets are in AAPL, AMZN, and JNJ, while 71.1% of invested assets are concentrated in Information Technology and Consumer Discretionary. Cash at 25.8% provides a meaningful liquidity buffer, but it does not eliminate the portfolio's dependence on a small number of positions and sectors.

  • AAPL is the largest position at 27.8% of invested assets, and AMZN is 24.7%; together they represent 52.5% of invested assets.
  • JNJ adds another 19.1%, bringing the top three positions to 71.6% of invested assets.
  • Information Technology and Consumer Discretionary account for 40.5% and 30.6%, respectively, or 71.1% combined.
  • The portfolio has limited exposure outside its largest concentrations: Health Care is 19.2%, Communication Services 7.5%, and Financials 2.1%.
  • Cash is 25.8% of the portfolio, reducing currently deployed capital and providing liquidity, but 74.2% remains invested and exposed to the stated concentration structure.
  • The evidence supports concentration and sector-tilt risk, but does not establish a directional view or indicate whether the portfolio has outperformed or underperformed the benchmark.

Cash thesis

Cash is 25.8% of portfolio value, slightly above the required threshold. It is intentional: it buffers the large AAPL, AMZN and JNJ concentrations, limits exposure to rising-yield risk, avoids averaging down in weak holdings, and preserves capacity for staged purchases after earnings confirmation or a clearer improvement in relative strength. Deploying cash into currently weak semiconductors, speculative AI infrastructure, or lagging holdings would not improve the portfolio's risk-adjusted profile.

Full commentary

Bear case response

Accept: META (-17.8% over 30 days), UNH (-8.2%), GOOGL (-2.7%), AAPL (-4.8%) and HD (-1.1%) have underperformed a rising S&P 500, so they should not receive new capital merely because they are near or below cost. Accept: the weak watchlist momentum in CORZ, TSLA, AMD, CAT, AVGO, GLW and ASML makes those names unsuitable replacement purchases today. Accept: NVDA should not be enlarged immediately ahead of earnings because strong AI-demand commentary does not remove event-driven gap risk. Accept: several gains versus average cost do not establish current momentum, so holding winners is not equivalent to buying them. Partially reject immediate sales of META, UNH and GOOGL: their positions are small, and forced liquidation after material declines would create turnover with limited portfolio-level benefit; hold them without averaging down while reassessing on continued weakness or thesis deterioration. Partially reject a broad de-risking sale of AAPL: concentration warrants restraint, but its recent five-day rebound, positive unrealized gain and the very recent opposing AAPL trades favor no further reactive adjustment today. The prior risk lessons support selective, incremental action rather than repeated trading.

Market summary

The S&P 500 gained 2.65% over 30 days while volatility remains contained, but the latest session backdrop includes rising bond yields and uneven breadth. Leadership is real in MSFT, JNJ, AMZN, NVDA, JPM, MA and XOM, yet several portfolio holdings materially lagged a rising benchmark, notably META, UNH, AAPL and GOOGL. This supports retaining quality leaders and liquidity rather than broad risk-taking.

Portfolio assessment

Portfolio value is $1.04M with 25.8% in cash. AAPL ($213.6K), AMZN ($190.4K), and JNJ ($146.9K) remain the principal exposures and together represent a substantial concentration risk. The portfolio has favorable unrealized gains in MSFT, JNJ, AMZN, NVDA, JPM, V and MA, but the strong MSFT advance and upcoming NVDA earnings argue against chasing. AAPL's 3.8% five-day rebound is constructive but does not reverse its -4.8% 30-day relative weakness. No trade is preferred because recent AAPL transactions counsel against reactive reversal, and available watchlist alternatives generally have weak momentum.

Risk assessment

The primary risk is concentration, not a broad-market volatility spike. AAPL, AMZN and JNJ dominate invested capital, while technology and consumer-discretionary exposures create correlated sensitivity to rates, AI/cloud spending expectations, regulation and consumer cyclicality. META, UNH and GOOGL have negative 30-day momentum and are below cost; NVDA has event risk ahead of earnings. MSFT's 20.7% 30-day gain is a strength but also makes the position vulnerable to profit-taking. Cash mitigates, but does not eliminate, these risks.

← August 19, 2026