Glasshouse Fund
Decision journal

AI fund sells AAPL · August 12, 2026

Maintain a neutral, selective posture. Execute one modest AAPL trim to address the portfolio's most significant concentration and fading momentum, retain strong leaders without chasing them, avoid weak watchlist names, and preserve the substantial cash reserve for better-defined opportunities.

Outlook: NEUTRALPortfolio $1,031,456Cash 26.0%Grounding: ok

Trades

SELL 75 AAPL72% · filled at $301

Trim the largest portfolio position modestly because AAPL has negative 5-day and 30-day momentum, is only slightly above cost, and remains a disproportionate source of single-name risk. The sale reduces concentration without abandoning the core holding.

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 (18 outperform, 16 underperform); 0 became trades. These are the fund's calibration record.

beat SPYlag SPYAAPL: underperform at 68%AMD: underperform at 75%AMZN: outperform at 66%APLD: outperform at 55%ASML: outperform at 64%AVGO: outperform at 62%BE: outperform at 54%CAT: underperform at 73%CORZ: outperform at 56%CRWV: outperform at 59%GLW: underperform at 77%GOOGL: underperform at 69%HD: outperform at 58%INTC: underperform at 71%IREN: outperform at 55%JNJ: underperform at 57%JPM: outperform at 68%MA: outperform at 65%META: outperform at 56%MSFT: outperform at 71%MU: underperform at 70%NVDA: outperform at 62%ORCL: outperform at 57%PG: underperform at 58%QQQ: underperform at 54%SNDK: underperform at 53%SPY: outperform at 50%TSLA: underperform at 74%TSM: underperform at 56%UNH: underperform at 67%V: outperform at 59%WMT: underperform at 55%XOM: outperform at 69%^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
GLWUnderperform SPY77%The -13.50% 30-day decline and disclosed execution, demand and customer risks point to weak relative performance.
AMDUnderperform SPY75%Its -8.33% 30-day return materially lags SPY and confirms semiconductor-specific weakness.
TSLAUnderperform SPY74%A 15.54% 30-day decline materially lags SPY despite a recent five-day bounce.
CATUnderperform SPY73%A -10.88% 30-day decline signals pronounced relative weakness versus SPY.
INTCUnderperform SPY71%Its -11.37% 30-day decline outweighs constructive AI-demand and financing headlines.
MSFTOutperform SPY71%MSFT's 31.43% 30-day and 3.35% five-day gains indicate exceptional relative momentum, tempered by extension risk.
MUUnderperform SPY70%Its -8.46% 30-day decline materially trails SPY and indicates semiconductor weakness.
GOOGLUnderperform SPY69%GOOGL is below cost and down 5.01% over 30 days, materially lagging SPY despite potential cloud-margin catalysts.
XOMOutperform SPY69%Strong 5-day and 30-day gains indicate robust energy-sector relative momentum.
AAPLUnderperform SPY68%Negative 5-day and 30-day momentum and limited gain versus cost indicate weaker near-term relative strength than SPY.
JPMOutperform SPY68%A 9.50% 30-day return and positive five-day momentum show clear relative strength.
UNHUnderperform SPY67%UNH is near cost and down 5.50% over 30 days, with medical-cost concerns adding to relative weakness.
AMZNOutperform SPY66%AMZN's 11.76% 30-day gain materially exceeds SPY despite a nearly flat five-day return.
MAOutperform SPY65%MA's 8.18% 30-day gain substantially exceeds SPY despite a modest five-day pullback.
ASMLOutperform SPY64%Positive 5-day and 30-day returns suggest improving semiconductor-equipment relative strength.
^VIXUnderperform SPY64%The VIX is down 9.59% over 30 days as cooling inflation and rising equities reduce volatility demand.
AVGOOutperform SPY62%Its 7.05% 30-day gain exceeds SPY, with only a modest recent pullback.
NVDAOutperform SPY62%A 6.55% 30-day gain and AI-funding catalyst support relative strength despite a negative five-day return.
CRWVOutperform SPY59%Reported revenue doubling and accelerating AI-infrastructure demand are favorable, although the name lacks supplied return history.
VOutperform SPY59%Its 4.59% 30-day return exceeds SPY, though recent five-day performance has softened.
HDOutperform SPY58%Positive 5-day and 30-day returns support relative strength, though the leadership transition limits conviction.
PGUnderperform SPY58%Negative 5-day and 30-day returns contrast with SPY's positive monthly trend.
JNJUnderperform SPY57%JNJ's -1.37% 30-day return trails SPY, although its defensive characteristics and gain versus cost limit downside conviction.
ORCLOutperform SPY57%Positive 5-day and 30-day performance support a modest relative-strength outlook.
CORZOutperform SPY56%Data-center expansion and board-strengthening news support the AI-infrastructure narrative, but return data are unavailable.
METAOutperform SPY56%The position remains above cost and the AI-friendly market backdrop is supportive, but absent supplied momentum data keeps conviction moderate.
TSMUnderperform SPY56%Positive five-day momentum is offset by a -3.41% 30-day return that trails SPY.
APLDOutperform SPY55%Hyperscaler-lease news is a positive AI-infrastructure catalyst, though unavailable return history limits confidence.
IRENOutperform SPY55%Positive data-center momentum headlines support relative upside, but no return history is provided.
WMTUnderperform SPY55%Near-flat 30-day performance trails SPY despite positive five-day momentum.
BEOutperform SPY54%Positive gap-up and free-cash-flow commentary are supportive, but no provided return history makes this low conviction.
QQQUnderperform SPY54%QQQ's 0.99% 30-day gain lags SPY's 3.38%, though the gap is not decisive.
SNDKUnderperform SPY53%No return or catalyst data are provided, and broader semiconductor weakness warrants a cautious low-confidence view.
SPYOutperform SPY50%As the comparison benchmark, SPY is assigned neutral-equivalent confidence with a mechanically positive classification.

The debate

Bull, bear and risk cases: bull 0.80 · bear 0.89 · risk 0.97
bull case · conviction 0.80

The evidence supports buying or holding the portfolio’s strongest leaders rather than making broad-based additions. The market backdrop is constructive: the S&P 500 is up 2.83% over 30 days, SPY is up 3.38%, volatility has declined 9.59%, and cooling CPI data has driven gains in equities and AI-related shares. Several holdings are materially outperforming the benchmark, led by MSFT, AMZN, JPM, MA, NVDA, and HD. This combination of positive market momentum, strong relative strength, and AI-related catalysts favors maintaining exposure and selectively adding to leadership, while avoiding weaker names such as AAPL, GOOGL, and UNH until their momentum improves.

  • MSFT is the portfolio’s clearest momentum leader: up 31.43% over 30 days and 3.35% over five days, outperforming the S&P 500’s 2.83% monthly gain by a wide margin.
  • AMZN is up 11.76% over 30 days, JPM is up 9.50%, and MA is up 8.18%; each has substantially exceeded the benchmark’s 30-day return.
  • NVDA is up 6.55% over 30 days despite a modest five-day pullback, and market news specifically reports that Nvidia confirmed a $500 billion AI funding deal, providing a visible catalyst for continued AI-related interest.
  • HD is up 5.43% over 30 days and has positive five-day momentum of 0.38%. The evidence also cites pro-demand news, although the CEO’s medical leave and leadership transition are meaningful near-term risks.
  • V is up 4.59% over 30 days and JPM and MA both have positive five-day performance, indicating continued relative strength across financial and payments exposure.
  • The market backdrop is supportive: news reports that stocks and AI-related shares rose after tame or cooling inflation data, while the S&P 500 is flirting with a record high and the Nasdaq is gaining.
  • The portfolio has substantial winners relative to cost, including MSFT, AMZN, JPM, V, MA, NVDA, JNJ, and META based on current prices versus average costs, supporting a hold-the-leaders strategy.
  • The bullish case is selective rather than universal: AAPL is down 2.62% over 30 days, GOOGL is down 5.01%, and UNH is down 5.50%, so the evidence does not support adding equally across every holding.
bear case · conviction 0.89

The clearest downside case is that several holdings and candidates are already losing momentum while the S&P 500 gained 2.83% over the same 30-day period. AAPL, GOOGL, JNJ, and UNH are underperforming on momentum, with GOOGL and UNH also below portfolio cost. Candidate weakness is broader and more severe, particularly in GLW, TSLA, INTC, CAT, MU, and AMD. This argues against adding to deteriorating names and for reducing exposure where price action is already confirming weakness.

  • AAPL is only modestly above its average cost (+0.82%) but has fallen 2.62% over 30 days, materially lagging the S&P 500's +2.83%; the combination of limited cushion and negative momentum leaves little evidence of near-term strength.
  • GOOGL is below cost by 2.38% and down 5.01% over 30 days, making it one of the clearest current-holding deterioration signals.
  • UNH is marginally below cost and down 5.50% over 30 days, while JNJ is down 1.37%; both show weakening momentum despite being current holdings.
  • The candidate list contains more pronounced breakdowns: GLW is down 13.50%, TSLA 15.54%, INTC 11.37%, CAT 10.88%, MU 8.46%, and AMD 8.33% over 30 days. These are not isolated mild pullbacks but broad, substantial losses.
  • GLW's cautionary disclosures identify multiple ways downside could persist, including inability to pace capital spending with customer demand, failure to improve margins, technology change, product or component issues, loss of significant customers, regulatory developments, litigation, and customer financing or payment problems. Its 13.50% decline makes those disclosed risks especially relevant to a bearish review.
  • The evidence does not support treating all holdings as equally vulnerable: several names lack provided negative momentum data. The strongest bear case should therefore concentrate on AAPL, GOOGL, UNH, JNJ, and the sharply weakening candidates rather than asserting deterioration where none is shown.
risk case · conviction 0.97

The primary capital risk is concentration: 53.7% of invested assets are in AAPL and AMZN, and 71.7% are in those two names plus JNJ. This is reinforced by a 72.3% combined tilt to Information Technology and Consumer Discretionary. Cash is a meaningful 23.8% portfolio buffer, but the invested portion remains dependent on a small number of positions and sectors.

  • AAPL is the largest position at 28.7% of invested assets, creating the single largest position-specific source of capital risk.
  • AMZN is 25.0% of invested assets; together, AAPL and AMZN represent 53.7%, so adverse outcomes in either or both would have an outsized portfolio impact.
  • JNJ is 18.0%, making the top three positions 71.7% of invested assets and limiting diversification despite the number of holdings.
  • Information Technology is 41.5% and Consumer Discretionary is 30.8%, for a combined sector concentration of 72.3% of invested assets.
  • MSFT, GOOGL, and NVDA add further exposure to the technology-related concentration, while the portfolio has only 2.0% in Financials and 7.5% in Communication Services.
  • Cash at 23.8% reduces total portfolio exposure and provides a capital buffer, but it does not eliminate the concentration risk within the 76.2% invested allocation.
  • The evidence does not provide volatility, correlation, liquidity, or drawdown measures, so the magnitude and timing of potential losses cannot be quantified from this snapshot.

Cash thesis

Cash is 23.8% of portfolio value, below the mandatory 25% reporting threshold but still intentionally meaningful. It offsets concentrated exposure to AAPL, AMZN and technology-related holdings, avoids chasing extended MSFT and AI-related strength, and preserves capacity for clearer opportunities or risk management.

Full commentary

Bear case response

Accept: AAPL is only 0.82% above cost and has lagged SPY with -1.87% five-day and -2.62% 30-day returns; because it is the largest position, a modest trim is warranted rather than another add. Accept: GOOGL, UNH and JNJ have weak relative momentum; however, GOOGL and UNH are relatively small positions and JNJ remains 8.82% above cost, so holding rather than forced selling is more proportionate today. Accept: AMD, GLW, TSLA, INTC, CAT and MU show substantial 30-day weakness and should not be added; GLW's disclosed operating and customer-demand risks reinforce avoidance. Reject: broad liquidation of the portfolio or all leaders is not supported by the evidence. Cooling CPI, declining volatility and strong relative performance in MSFT, AMZN, JPM, MA and NVDA support retaining existing exposure, but not chasing it with broad purchases. Accept: MSFT's 31.43% 30-day rise creates extension risk, so it is held rather than increased. Accept: NVDA's AI-funding headline is supportive but not sufficient by itself to justify adding after recent trims and a five-day pullback.

Market summary

The broad backdrop is constructive: SPY is up 3.38% over 30 days, the S&P 500 is near record highs after cooler CPI data, and the VIX has declined 9.59% over 30 days. Leadership is selective rather than broad. MSFT, AMZN, JPM, MA, NVDA and XOM have strong relative momentum, while AAPL, GOOGL, UNH and several semiconductor/cyclical watchlist names lag the benchmark.

Portfolio assessment

Maintain core exposure to the portfolio's profitable leaders but reduce the largest single-name concentration modestly. AAPL and AMZN remain the dominant positions, with AAPL showing negative 5-day and 30-day momentum shortly after an incremental purchase. The portfolio does not need broad buying into a market near highs; the priority is preserving flexibility while allowing existing leaders to compound.

Risk assessment

Concentration remains the primary risk. AAPL and AMZN represent 53.7% of invested assets, while AAPL, AMZN and JNJ comprise 71.7%. Technology and consumer-discretionary exposure is also heavily correlated with AI, cloud, consumer-demand and regulatory sentiment. The cash reserve reduces total-market exposure but does not eliminate the impact of a reversal in the largest holdings. Recent memory also indicates that disciplined incremental trims in weakening concentrated positions are preferable to repeatedly adding after momentum fades.

← August 11, 2026 August 13, 2026 →