Mag 7, Actively Managed

I'm tilting away from an equal-weight (14.3% each) approach based on clear momentum and fundamental divergence within the group this week. Microsoft just posted a blockbuster earnings report with Azure revenue crossing $100 billion annually and 43% growth, sparking its biggest three-day rally in 26 years and reversing its 2026 losses - the data confirms this isn't fading (still +3.8% this week, +31% this month, only 6% off its 52-week high), so I'm raising it to 30%, the largest allowable position given the 35% cap. Nvidia gets bumped to 20% for its steady momentum (+5.3% this week) combined with the lowest volatility in the group, making it an efficient way to keep AI exposure. Apple, a stable performer with moderate gains and below-average volatility, stays near equal-weight at 15%. On the other side, Meta is cut sharply to 5%: it missed earnings expectations, its free cash flow plunged as AI spending ballooned, and the stock sits nearly 24% below its 52-week high with negative returns over the last one, three, and six months. Alphabet drops to 8% as it's down over the past week and down more than 10% over three months with no offsetting catalyst in the data. Amazon is trimmed modestly to 18% - its 6-month trend is strong, but the most recent week turned negative and its volatility is the second-highest in the group, so I'm not increasing exposure further. Tesla remains excluded entirely: it's the weakest name in the group on every timeframe, down double digits over the last month and sitting nearly a third below its 52-week high, so there's no case for owning it right now. This leaves roughly 4% in cash as a small buffer, well within the 0% minimum-cash rule.

This week's Mag7 earnings season created a clear split. Microsoft and Amazon delivered blowout results with accelerating cloud growth (Azure +43%, AWS +37%, its fastest pace in 18 quarters) and strong forward guidance, sparking the biggest rallies in the group. Alphabet's Google Cloud grew a stunning 82%, also justifying a moderate overweight. Apple, by contrast, warned that memory-chip supply constraints will worsen next quarter and missed on services, triggering its worst four-day stretch in over a year — I'm trimming it from roughly 30% of the portfolio down to 15%, still a meaningful position but no longer the largest bet given the cloudy near-term outlook. Meta beat on revenue but missed badly on earnings as free cash flow collapsed under heavy AI capex, so I'm trimming it modestly to 10%. Nvidia's story didn't change much this week, so I'm leaving it close to where it was. Tesla continues to show negative momentum across every timeframe (down double digits over 1-, 3-, and 6-months) and stays excluded. Versus simple equal-weighting (14.3% each), this positioning leans into companies that just proved their AI spending is paying off in hard revenue numbers, while reducing exposure to the two names facing genuine near-term headwinds (supply constraints for Apple, cash-flow burn for Meta). About 4% is left in cash as a small buffer given the mixed signals from Meta and Apple.

Equal-weighting all seven would put each stock at ~14.3%, but the seven are not equally strong right now, so I'm tilting hard toward the momentum leader and away from the clear laggard. Apple is the standout: it just hit a fresh all-time high near $337, is up over 20% this year, and has a strong history of beating estimates heading into its July 30 earnings. I'm pushing it to the 35% cap (from ~32%), the maximum allowed, to capture that momentum while accepting earnings-day risk. Nvidia gets a small bump to 16% as the steadiest large-cap performer of the group with no major red flags in the data. On the other end, Tesla is cut to zero. Its Q2 report showed operating margin collapsing to 1.4% and free cash flow turning negative, and the stock just had its worst weekly performance since 2022 with a roughly 15-18% single-day plunge on the earnings miss. Combined with the highest volatility in the group (64% annualized) and a stretched valuation, there's no case for holding it through the next week. Alphabet is trimmed to 11%: it already reported earnings alongside Tesla and by all accounts got a lukewarm market reaction, and its 1-week (-7.2%) and 1-month (-5.0%) returns are among the weakest in the group. Amazon is trimmed slightly to 12% and Meta held near 14%, both facing earnings this week (July 29-30) with elevated volatility and heavy AI capex scrutiny — worth keeping meaningful exposure given their historical beat rates, but not adding into the uncertainty. Microsoft stays underweight at 8%: despite a recent bounce, it remains one of the worst Magnificent Seven performers this year, down over 27% from its 52-week high on 'SaaSPocalypse' fears and rising capex, with an earnings report also due July 29. The leftover ~4% is held in cash rather than force-fed into a group whose average trailing returns are lagging the S&P 500, reflecting a modestly cautious stance heading into a week packed with four Magnificent Seven earnings reports.

Equal-weighting the Mag 7 would put each name at 14.3%, but the group's momentum and news backdrop is far from uniform, so I'm tilting away from that baseline. AAPL gets the biggest overweight (30% vs 25.9% now): it's the strongest performer in the group across every timeframe (1w +4.1%, 1m +10%, 3m +20%, 6m +28.5%), sits just 1.8% off its 52-week high, and is described as 'the best-performing member of the Mag 7 ahead of NVIDIA and Alphabet,' with a fresh record high driven partly by recent price hikes. NVDA stays near-equal-to-slightly-overweight (16%): despite a short-term dip, the AI infrastructure capex cycle from hyperscalers remains intact, with industry executives calling demand 'almost unlimited.' MSFT moves up sharply from a tiny 3.1% to 8%: it just posted the best 1-week (+3.3%) and 1-month (+4.8%) returns in the group, suggesting the worst of its 6-month drawdown (-13%) may be stabilizing, though it's still 26% off highs so I'm not going overweight. GOOGL gets trimmed from 17.4% to 14% (roughly equal-weight) given weak recent momentum (-3.4% 1w, -5.7% 1m) despite decent 3-6m trends. AMZN rises modestly from 10% to 13%, reflecting flat-but-stable performance with no major red flags. META is cut significantly from a 24.4% overweight to 15%: its momentum has turned negative over the past week/3 months, it carries the highest volatility in the group (49% annualized), and its pivot into selling excess AI cloud capacity has stirred mixed signals about industry overcapacity. TSLA is left near its current small 4% weight: it has the weakest returns across every timeframe, the highest volatility (55%), a very rich valuation, and reports earnings today after the close with options pricing a ~7.6% swing — a binary risk best kept small rather than added to or eliminated entirely, since the outcome could go either way and we won't get to react again for a week. This concentrates the portfolio in the two strongest, most durable growth stories (Apple, Nvidia) while de-risking the more volatile or momentum-losing names (Meta, Tesla), a deliberate departure from simple equal-weighting.

Compared to a simple equal-weight (14.3% each), this week's tilt leans into the two names showing the broadest, most consistent strength and trims the two showing clear fatigue, while keeping the book fully invested since several Mag7 names still show real momentum rather than group-wide exhaustion. META gets the largest increase (16%→25%). It just posted its best week since early 2024 after unveiling 'Meta Compute,' a new AI cloud business, with shares up double digits on the week and analysts flagging real EPS upside from monetizing compute capacity. Fundamentals back the move: the business is still growing revenue at a rapid clip with strong margins and cash generation. Its earnings report isn't until July 29, safely after our next decision, so this week's holding period avoids that binary risk. AAPL also gets bumped up (21.7%→25%). It shows strong, consistent returns across every timeframe (1-week, 1-month, 3-month, and 6-month all positive) and sits just below its 52-week high, the steadiest tape in the group. NVDA is increased meaningfully (8.1%→15%) on the back of a strong week (+7.6%) driven by eased export restrictions to the UAE and continued hyperscaler capex commitments, while trading at a historically low forward P/E for the name, suggesting the recent laggard status versus peers may be turning. GOOGL is trimmed from an overweight 27.3% to 18%. It's still a solid business, but it was our largest position (near the max cap) and just posted a negative week (-2.1%) with an earnings report expected around July 22 — inside our holding window — so we're right-sizing the concentration ahead of that event risk. AMZN is trimmed to 10%. Fundamentals remain fine, but price momentum has been essentially flat over the past 1-6 months, so it doesn't earn an overweight versus cash we can redeploy to stronger names. TSLA is cut most aggressively (9.9%→4%). Momentum is negative on the week and over six months, the stock trades at a steep valuation versus weak profit margins, and it also has an earnings report around July 22 that falls within this holding period, adding event risk we'd rather not carry at a large size. MSFT is left roughly unchanged (~3%), reflecting its persistent underperformance — down materially over six months and off its 52-week high — as heavy AI capex has pressured free cash flow and investor sentiment; we keep only a small stake rather than dropping it entirely since the underlying cloud/AI business (Azure growth, large backlog) still has credible upside if capex worries ease.