Mega-Cap Tech Drubbing Deepens as NVDA Exits on Chip-Price Headwind

Monday, June 22, 2026

The Monday session delivered a selective but meaningful repricing across technology. Mega-cap names that had been priced for AI earnings acceleration came under renewed pressure: AMZN fell 4.02%, AVGO dropped 4.04%, META declined 2.03%, and MSFT shed 2.41% — each breaking back below their respective 20-day moving averages and confirming that June 19's tech rally was narrower in breadth than the headline QQQ and XLK recovery suggested.

QQQ closed at $735.85, down 0.64%, but held above SMA20 ($727.94). That resilience reflects index composition more than breadth — the names doing the damage today carry significant QQQ weight but were partially offset by Apple, which CNBC flagged as a notable outlier on the day. XLK, the pure-play technology sector ETF held at 10% in the portfolio, closed flat (0.00%) and retained its SMA20 reclamation from June 19. Its 60-day momentum of +14.79% remains the highest reading across the entire candidate universe, even as individual constituents inside it are breaking down.

SPY closed at $743.89, down 0.38%, now sitting $3.25 below its own SMA20 ($747.14). The broad market is technically softer than the tech ETF level suggests. VIX at 17.41 is elevated enough to signal residual uncertainty without indicating panic.

NVDA Exited: A Thesis in Progress Meets a New Headwind

NVDA closed at $208.55, down 1.01%, still $2.70 below SMA20 ($211.25). The position has been held at minimum conviction size (5%) since June 16 with a single explicit criterion for rebuilding: a clean SMA20 reclamation confirmed for one full session. That criterion was not met for the fourth consecutive session since the position was initiated.

More significantly, a new bearish catalyst emerged independently of the rate-discount story: Kalshi prediction market traders are now explicitly betting that chip prices will decline in the near term. This is a direct headwind to NVDA's earnings growth narrative — if chip pricing is expected to compress, the unit economics of NVDA's data center revenue face margin pressure regardless of whether hyperscaler capex commitments remain intact in aggregate volume terms. The prediction market signal is not a confirmed fundamental development, but it represents real capital positioning against the thesis that justified holding NVDA at 5%.

With the technical trigger unmet and a new specific fundamental headwind registering in forward markets, the NVDA position is exited. Cash moves from 40% to 45%.

Iranian Oil Authorization Deflates the Hormuz Risk Premium

The Treasury Department authorized Iranian crude oil sales through August — a material development that directly unwinds the supply-restriction premium that had supported energy prices during the period of heightened Hormuz friction. Oil prices fell on the announcement. XLE managed a small positive session (+0.43%) suggesting the market had partially priced this outcome, but the directional read is unambiguous: the geopolitical energy tail risk that has provided a floor for energy pricing is diminishing.

XLE remains 5.26% below SMA20 ($57.00) with 60-day momentum of -6.62%. The sector trend is negative and the Iranian authorization removes rather than creates a catalyst to own it. No energy position is held, and the incoming evidence does not support initiating one.

AI Infrastructure: Commitment Is Real, Monetization Is Not Yet

The Chevron–Microsoft natural gas data center deal in Texas is the structurally important story underneath the surface selloff. Hyperscalers are signing direct agreements with fossil fuel producers to guarantee dispatchable, scalable power for AI compute — the infrastructure commitment is now measured in gigawatts and is not subject to quarterly repricing. This confirms that the AI capex cycle has crossed from software allocation into physical infrastructure in a way that creates durable demand across energy, construction, and semiconductor supply chains.

Yet simultaneously, today's mega-cap selloff and the Kalshi chip-price prediction market signal that AI spending expectations are being front-loaded into price rather than earnings at the individual equity level. The market is distinguishing between AI commitment — locked in at the infrastructure and governance level — and AI monetization — still uncertain at the application layer where revenue-per-user economics are unproven at scale.

QQQ and XLK remain in the portfolio precisely because they provide diversified exposure to the AI earnings thesis without requiring a single-name bet on monetization timing. Individual mega-cap names — AMZN, AVGO, META, MSFT — are not held because their technicals have deteriorated and their specific risk-reward profiles have worsened relative to the diversified ETF alternatives.

Political Context: UK Leadership Transition

Keir Starmer's replacement by Andy Burnham as UK Prime Minister introduces a degree of political uncertainty in a G7 economy. CNBC flagged the transition as relevant to markets, though the direct portfolio impact is limited given the domestic equity focus of this book. UK political transitions typically generate short-term sterling and gilt volatility; none of the four portfolio positions carry meaningful UK revenue exposure.

Institutional Positioning Signals

Bridgewater's 13F (filed May 15) shows broad index exposure — SPY at 12.7% and IVV at 10.7% — with selective AI single-stock exposure including NVDA at 3.7%. Bridgewater's positioning is structurally consistent with the view that AI upside is best captured through broad market exposure rather than concentrated single-name bets, a thesis that today's mega-cap drubbing validates in the short term.

Pershing Square holds MSFT at 15.3% of its portfolio. MSFT was down 2.41% today and is now -9.85% on a 20-day basis. Bill Ackman's concentrated MSFT bet is underwater on the month — a useful reference point for how much concentrated single-name tech risk is embedded in smart-money books, and a directional argument for the portfolio's ETF-based tech exposure over individual large-cap software names.

Berkshire Hathaway was essentially flat today (-0.05% in BRK-B) despite the broad selloff, trading at $489.21 and holding above SMA20 ($483.98). Buffett's portfolio — clustered in AAPL, AXP, KO, BAC, and CVX — provides a working benchmark for what genuine defensive resilience looks like when high-multiple growth is repricing. Berkshire's flat performance while MSFT and AMZN dropped 2–4% is a datapoint about the durability of value and consumer-franchise positioning in the current rate environment.

Michael Burry's large PLTR concentration (66% of Scion AUM as of the November 2025 13F) is outside the candidate universe and not actionable for this portfolio, but the filing confirms Burry is making a single large concentrated bet on AI data infrastructure rather than diversified AI exposure — a positioning approach materially different from this portfolio's framework.

Current Portfolio Positioning

The portfolio holds four positions totaling 55% of NAV with 45% in cash:

JPM (15%) — The rate-hike thesis is the most directly reinforced position in the book. JPM gained 2.19% today, now 6.54% above SMA20. NIM expansion remains the primary driver and is gaining further support from a rate environment where 50%+ hike probability has persisted for two weeks. No change.

LLY (15%) — Up 0.53%, recovering toward SMA20 ($1,110.86). The GLP-1 structural demand thesis is mechanically insulated from the rate environment. Momentum60 of +10.55% is the second highest in the universe. Technical recovery is in progress but not confirmed at SMA20.

QQQ (15%) — Held above SMA20 despite the mega-cap selloff. The buffer has narrowed. This is a monitored hold; the position will be reassessed immediately if $727.94 is breached on volume.

XLK (10%) — Flat on the day, SMA20 intact, highest 60-day momentum in the universe. Held at current weight; rate uncertainty and mega-cap pressure cap further expansion.

The 45% cash is intentional and not a residual. The rate-hike probability above 50% remains the dominant governor of this portfolio's risk budget. Forced aggression in an environment where NVDA is generating specific chip-price headwinds and MSFT is -9.85% on the month is a risk that 45% cash explicitly hedges against.

What Would Change the Thesis

Upside triggers: NVDA reclaiming SMA20 ($211.25) on a session where the Kalshi chip-price prediction market reverses re-opens the 5% semiconductor entry at minimum conviction size. QQQ sustaining above SMA20 through the week while mega-cap selling pressure stabilizes would validate the tech reclamation sufficiently to consider rebuilding toward 20%. A Fed statement that meaningfully reduces rate-hike probability would allow the portfolio to expand its risk-on weight more broadly.

Downside triggers: QQQ breaking below SMA20 ($727.94) on volume would signal the June 19 reclamation was a false recovery, warranting trimming of both QQQ (from 15% toward 10%) and XLK (from 10% toward 5%) and a further cash build toward 55–60%. A Fed rate hike announcement ahead of the September meeting would compress tech multiples further — JPM would likely be the survivor while QQQ and XLK face reduction. A second consecutive LLY close below SMA20 as the recovery stalls would prompt a sizing review of that position.