Saturday, June 27, 2026

Your Saturday morning market coffee. The week the AI supercycle sent its invoice: South Korea's KOSPI triggered a circuit breaker on Tuesday after plunging 9.99% as Samsung and SK Hynix each fell more than 12% — the catalyst a perfect three-way collision of MSCI re-excluding South Korea from its Developed Markets watchlist (removing the passive-inflow premium that foreign investors had pre-bought), leveraged single-stock chip ETF regulatory concerns, and forced retail liquidations amplified by the Fed's hawkish June 17 dot plot — dragging the Nasdaq down 2.21% and the Philadelphia Semiconductor Index down approximately 7% in what became the worst single session for global chip stocks since the yen-carry-trade unwind of August 2024; only for Micron Technology to report Q3 FY2026 earnings after Wednesday's close that were so exceptional they bordered on surreal — EPS of $25.11 against a $20.28 estimate (a 24% beat), revenue of $41.46 billion against a $35.25 billion estimate, gross margins of 84.9% (the highest in the company's history), and Q4 guidance of $50 billion in revenue and $31 EPS — prompting the Nikkei 225 to surge 4.61% on Thursday in what Tokyo traders described as the most concentrated AI-hardware demand confirmation in a single data release in years; but that same Micron quarter revealed the mechanism by which AI infrastructure wealth is transferred downstream to consumers: Apple announced price increases of $200 or more on every Mac and iPad model on Thursday (MacBook Air: $1,099 → $1,299; iPad Pro: $999 → $1,199), Microsoft announced Xbox price increases of $100–$150 beginning August 1, and analysts projected iPhone hikes of $50–$200+ in September — all explicitly attributed to AI-driven memory chip cost surges in which Micron's 84.9% gross margin is the accounting identity of Apple's $200 price increase; the same Thursday that Apple and Microsoft raised prices saw the Bureau of Economic Analysis release May PCE data showing headline inflation at 4.1% and core at 3.4% — the highest readings since April 2023 and October 2023 respectively — but the market's reaction was nearly flat (S&P -0.01%) because traders understood that May PCE is a photograph of the Iran war's oil prices, not a window into the current sub-$71 world where WTI has completed three consecutive weekly declines from approximately $80–81 at the Iran MOU signing (the level was ~$84–85 the week before) to below $71 today; and on Friday, a New York Times report that OpenAI has decided to delay its IPO until 2027 — with CEO Sam Altman refusing to consider any valuation below $1 trillion and advisory teams presenting a binary choice between a 2027 listing at $1T or a 2026 listing at a discount — sent SoftBank Group shares down more than 12% (their largest single-day decline since August 2024) while simultaneously confirming the lesson SpaceX had already taught: the largest IPO in history ($225.64 ATH on June 16) now trades at $153, and the largest remaining planned IPO in history won't be priced until the first has stabilized; with the Nasdaq-100's official confirmation of SpaceX's July 7 inclusion arriving on the same Friday, the net result of a five-day week was S&P 500 down 1.95% to 7,354.02, Nasdaq down 4.60% to 25,297.62, the Dow — the safety-rotation winner of the week — up 0.60% to 51,876.11, gold below $4,000 for the first time since November 2025, Bitcoin at $59,419 (-7% for the week), and Switzerland's talks producing an Iran-deal roadmap that was the week's only unambiguous positive — a market in which AI is simultaneously the biggest earnings story, the biggest inflation story, and the biggest IPO story, with all three pointing in different directions at once.

Week in Review

The Numbers

Note: Full five trading days (Mon June 22 – Fri June 26). Opening prices on Monday June 22 are estimated from the prior Thursday June 18 close (the last US session before Juneteenth); Friday June 26 closes are confirmed.

The Story Arc

Monday June 22 opened with the week's only unambiguously positive macro development: the US–Iran Switzerland talks, which the prior week's cancellation scare had made uncertain, had in fact proceeded on June 21–22, producing the most substantive progress since the June 17 MOU signing. Mediators Qatar and Pakistan announced that the two sides had agreed on "a roadmap towards reaching a final deal within 60 days," Iran had agreed to invite IAEA inspectors back into the country — the first such agreement since Iran expelled inspectors in July 2025 — and a Lebanon deconfliction plan had been added to the framework. Three working groups were established: one on nuclear terms, one on sanctions, and one on oversight. Markets opened the week modestly positive on the Iran-roadmap news; WTI continued its descent below $73 as the deal-completion probability rose; the Nikkei 225 hit its weekly high of approximately 72,831 on Monday, a new all-time high at the time of trading.

Tuesday June 23 was the week's most violent session, and one of the more dramatic single-day events in global equity markets in 2026. Three forces collided simultaneously in Seoul. First, MSCI announced it would again exclude South Korea from its Developed Markets watchlist — the latest in a series of annual exclusions (approximately the 12th overall setback since South Korea was first placed on the DM watchlist in 2008) — removing the passive-inflow catalyst that foreign institutional investors had been front-running for months. Second, South Korean financial regulators raised concerns about leveraged single-stock ETFs tied to Samsung Electronics and SK Hynix that had only been approved the prior month. Third, the broader repricing of AI stocks in the context of the Fed's hawkish June 17 meeting (nine FOMC members projecting a rate hike, forward guidance dropped) prompted a wave of forced liquidations among retail investors who had used borrowed money to buy chip stocks at elevated valuations. The combined effect: the KOSPI plunged 9.99% — its fifth-largest single-day fall on record — triggering a 20-minute circuit-breaker halt. Samsung Electronics fell more than 12%; SK Hynix fell more than 12%. More than $2.5 billion in foreign capital left the Korean market in a single session. The contagion was immediate and global: the Nikkei reversed from Monday's high, falling 3.55% on Tuesday as chip names across Asia were repriced; the Philadelphia Semiconductor Index fell approximately 7% in the US session; the Nasdaq Composite closed down 2.21%; the S&P 500 closed down 1.44%. Bank of America had also published a note on Monday June 22 highlighting the September rate-hike risk, adding a US-rate-policy overhang to the Korean-triggered selloff.

Tuesday June 23 saw FedEx and Carnival earnings, followed on Wednesday June 24 by Micron after the close. FedEx Q4 FY2026: EPS $6.31 vs $5.92 estimate (beat), revenue $25.0 billion vs $24.01 billion estimate (beat), revenue +12.5% year-over-year — but shares fell 3.63% in the regular session and another 6.16% after hours, as investors focused on the company's FY2027 guidance and cost pressures stemming from the FedEx Freight spin-off transition. The Network 2.0 cost-savings initiative ($1 billion targeted for calendar 2026) is on track, but the transition year creates near-term EPS uncertainty that the market penalized. Carnival Q2 2026: adjusted EPS $0.41 vs $0.34 estimate (21% beat), revenue $6.66 billion, adjusted net income a record $569 million — yet shares declined in a classic sell-the-news reaction; customer deposits of $9.0 billion represent an all-time high and confirm that leisure travel demand is robust despite inflation headwinds. After the close, Micron delivered a quarter that restructured the AI hardware thesis in a single data release: EPS $25.11 vs $20.28 estimate (+24%); revenue $41.46 billion vs $35.25 billion estimate; gross margin 84.9% (a company record); Q4 FY2026 guidance of $50 billion in revenue, approximately 86% gross margin, and $31 EPS — guidance itself exceeding what had been the consensus for the full Q4 by a wide margin. Data-center revenue exceeded $25 billion in the quarter, representing an annualized run rate above $100 billion for that segment alone. An Anthropic strategic supply and investment agreement announced on June 22 added a named, multi-year HBM demand anchor to the AI buyer list. Micron shares jumped 13.1–14.6% in after-hours trading. The Nikkei 225, tracking the US AI news in overnight trading, was already positioned for Thursday's surge.

Thursday June 25 was the week's most data-dense session. The Nikkei 225 opened to Micron's after-hours surge and climbed 4.61%, closing at approximately 72,366 — one of the largest single-day gains in the Japanese benchmark in 2026. The euphoria, however, collided with two morning announcements in the US that complicated the AI narrative. Apple disclosed price hikes across its full Mac and iPad lineup: the MacBook Air rose from $1,099 to $1,299 (+$200, +18%); the MacBook Pro from $1,699 to $1,999 (+$300, +18%); the iPad Air from $599 to $749 (+$150, +25%); the iPad Pro from $999 to $1,199 (+$200, +20%). Apple explicitly attributed the increases to AI-driven memory chip cost surges. Microsoft simultaneously disclosed Xbox console price increases: the 512GB model +$100 and the 1TB model +$150, effective August 1, stating that "memory prices have increased by more than 2.5× and we expect another doubling by the fall of 2027." Apple fell 6.13% on the session; Microsoft fell approximately 3.5%. The Magnificent 7 stocks began their most significant stretch of underperformance of the year. At 8:30 a.m., the Bureau of Economic Analysis released May 2026 PCE data: headline PCE +4.1% year-over-year (up from 3.8% in April), the highest since April 2023; core PCE +3.4% (up from 3.3%), the highest since October 2023. Both readings matched or slightly exceeded expectations. The market's reaction was essentially flat (S&P 500: -0.01% to 7,357.49), which tells the entire story: traders know May PCE is measuring the energy prices of the Iran war period (WTI peaked near $106/barrel in early May 2026 and averaged approximately $100–102/barrel for the month when the Hormuz blockade was near-total), not the $71/barrel present. Darden Restaurants reported Q4 FY2026: adjusted EPS $3.66 vs $3.64 estimate; same-store sales +4.6% vs 4.1% consensus; revenue $3.72 billion (+13.7% year-over-year, including a 53rd-week benefit). Full-year FY2026 sales exceeded $13 billion for the first time. Shares fell approximately 1% — the consumer recovery story acknowledged, the valuation held fair.

Friday June 26 delivered the week's final blow to AI sentiment, and one of the more consequential IPO-market signals since SPCX's own listing. The New York Times reported that OpenAI's advisory team has presented management with two options: list in 2026 at a valuation below $1 trillion, or wait until 2027 and target a $1 trillion listing. CEO Sam Altman has reportedly made clear that a below-$1-trillion proposal is unacceptable. The implicit context: SpaceX, the only comparable AI-era IPO, has fallen from its $225.64 intraday high on June 16 to approximately $153 by June 26 — a 32% decline from the ATH, driven by the same AI multiple decompression that has plagued the Nasdaq all week. SoftBank Group, the single largest private investor in OpenAI via its Vision Fund, fell more than 12% — its worst day since August 2024. Against that backdrop, Nasdaq officially confirmed on Friday that SpaceX (SPCX) will join the Nasdaq-100 Index on July 7, 2026, with a weighting of under 1% — a lower weight than the original estimates of 6–8% that were based on SpaceX's $225 ATH valuation; at $153, SPCX's market cap and index weight are both materially reduced. FTSE Russell simultaneously added SpaceX to its US equity indexes effective after Friday's close as part of the semi-annual reconstitution (the first since June 2025) — a mechanical forced-buy event for Russell 1000 and Russell 3000 trackers. The broader market on Friday: S&P 500 -0.05% to 7,354.02; Nasdaq -0.24% to 25,297.62; Dow -0.09% to 51,876.11. WTI crude fell below $71 per barrel — the third consecutive weekly decline, down more than $6 from the June 18 close of $77.33.

The week's narrative: Micron's quarter was historic, and it explains everything: AI is consuming memory at prices that produce 84.9% gross margins for the maker, $200 price hikes for the consumer, and a $41.5 billion quarterly revenue number that makes Micron, briefly, one of the highest-revenue companies in the world on a run-rate basis. But the market's reaction to Micron's perfect quarter — jumping 14.6% after hours, then closing the week roughly flat — encapsulates the week: even a flawless earnings report cannot sustain a rally in a market where hot PCE (4.1% headline), KOSPI circuit breakers, Apple price hike sticker shock, and OpenAI's IPO-delay signal are competing for the sentiment bandwidth. The Dow's +0.60% vs. Nasdaq's -4.60% is a 5.2 percentage point spread — the largest single-week rotation from growth to value/defense since March 2026.

Biggest Movers

Winners

Losers

Market Scoreboard

Weekly Index Performance

Note: Russell 2000 figure is approximate; FTSE Russell reconstitution took effect after the Friday June 26 close, adding SpaceX and rebalancing the index for the first time since June 2025 (this June 2026 is the inaugural semi-annual reconstitution).

The Dow's +0.60% vs. Nasdaq's -4.60% — a 5.2 percentage point divergence — is the most significant single-week growth-to-value rotation of 2026. The Dow holds industrials (Caterpillar, Boeing, Honeywell), financials (Goldman, JPMorgan), energy names, and healthcare — none of which are exposed to AI memory cost inflation the way consumer tech hardware is. Apple's -6.13% Thursday contributed approximately 100–130 Dow points of drag (Goldman Sachs holds the Dow's largest weighting at ~12.5% under the index's price-weighting methodology, while Apple's lower share price gives it a more modest weight), while Microsoft's ~3.5% decline added further drag. The Dow's constituents are beneficiaries of the same AI capex boom through power demand (Honeywell) and adjacent power-demand names like GE Vernova (S&P 500, not a Dow component), industrial automation (Caterpillar for data center construction), and financial services (Goldman's AI M&A advisory pipeline) without being directly exposed to the DRAM/NAND cost spiral that is inflating consumer hardware prices.

The Nasdaq's -4.60% for the week is the index's worst single-week performance since early February 2026, when rate fears and the then-early stages of the Hormuz disruption sent the index lower. The composition of this week's decline is different: February was macro-driven (rates + geopolitics); this week is structure-driven (MSCI Korea exclusion forced liquidations, AI cost-inflation consumer impact, OpenAI IPO delay repricing the AI narrative). The structural interpretation: markets are not backing away from the AI theme, they are repricing where in the AI stack the excess returns accrue — from consumer devices (Apple, Microsoft) toward the infrastructure layer (Micron, but even Micron couldn't sustain its Thursday AH gains through Friday's risk-off).

The Nikkei's wild week — Monday ATH at ~72,831, Tuesday -3.55% (Korea contagion), Thursday +4.61% (Micron euphoria), Friday -4.15% (global tech selloff) — ended approximately -2.7% from the prior Japan close of 71,250 (June 19). The sogo shosha and industrial finance names absorbed the volatility better than Japanese chip equipment makers, which tracked the Philly SOX more closely.

Commodities & Rates

WTI at sub-$71 completing its third straight weekly decline from approximately $80–81 at the Iran MOU signing (June 17; the level was ~$84–85 the week before the signing) is the week's most important macro variable, though not the most discussed. From a cumulative perspective, WTI has now declined approximately $13–14 per barrel in nine trading sessions following the Hormuz physical reopening. At the current pace, the June average WTI will be approximately $73–76 — versus May's estimated average of $100–102 during the height of the Hormuz disruption. The mechanical PCE implication of that differential is significant (see the Fun Section below). The Switzerland roadmap advancing — IAEA inspections agreed, nuclear and sanctions working groups established — continues to price WTI toward the final-deal scenario ($68–74) rather than the MOU-collapse scenario ($82–90).

Gold's breach of $4,000 on Thursday June 25 — the first time since November 2025 that spot gold traded below that level — is significant not because $4,000 is a fundamental threshold but because it represents the unwinding of two simultaneous bid sources that had driven gold to approximately $4,362 on June 9 (its Q2 2026 high; the 2026 all-time high was ~$5,589 in late January): the safe-haven bid from the Hormuz closure and the inflation-hedge bid from escalating PCE fears. Both are now deflating simultaneously. The irony: gold fell below $4,000 on the same day PCE printed at its highest since April 2023. The market's interpretation — correctly — is that 4.1% PCE is backward-looking oil war data, not a signal for sustained inflation requiring a hard-asset hedge at $4,300. Gold recovered from its Thursday intraday lows to approximately $4,080 by the end of the week but the technical damage (break of the $4,000 support that had held since November) shifts the momentum.

The 10-year Treasury fell approximately 8 basis points on the week to ~4.38% — a counter-intuitive reaction to hot PCE that reflects the bond market's view that May PCE is backward-looking Iran-war oil data. The move was driven by: (1) KOSPI-triggered risk-off buying of Treasuries on Tuesday; (2) bond market's understanding that the May PCE is backward-looking; (3) modest flight-to-safety from the tech selloff. The 30-year at ~4.86% is an important level to watch — as it approaches 5%, that represents the threshold at which mortgage rates and corporate debt costs become materially restrictive, and the proximity to the 5% ceiling that triggered the May bond-vigilante reaction (see the May 16 fun section) bears watching.

Earnings Recap

The Earnings That Rewrote the AI Hardware Narrative: Micron Q3 FY2026

Micron's Q3 FY2026 is the most consequential earnings report since Nvidia's May Q1 FY2026 print in terms of its structural implications — but the implications are different and more ambiguous. Nvidia's quarter confirmed AI demand; Micron's quarter confirms AI demand and simultaneously explains why every device that uses memory is about to get more expensive. The 84.9% gross margin — achieved in a quarter where data-center memory demand from AI training and inference workloads entirely supplanted consumer DRAM demand as the pricing anchor — means that Micron is capturing an extraordinary rent from an infrastructure bottleneck. Every percentage point of Micron's gross margin compression is a percentage point of cost relief for Apple, Dell, HP, and every server manufacturer. At 84.9%, none of that relief is imminent.

The Q4 guidance of $50 billion in revenue is the number that stops the analysis. Annualized at $200 billion, it would make Micron one of the highest-revenue technology companies in the world — comparable to Meta in total sales. That is the AI HBM supercycle's signature: memory, historically a commodity with thin margins and cyclical troughs, is temporarily pricing like a monopoly good because HBM3E and HBM4 production capacity cannot be built at the speed AI demand is growing.

The stock's failure to sustain the +14.6% after-hours reaction is itself a signal. Investors celebrated the quarter but sold into the strength because: (1) the hot PCE data on Thursday morning raised real rate concerns; (2) Apple's price hike announcement on the same morning framed Micron's margins as consumer-price inflation in disguise; and (3) the broader tech-selloff context (Philly SOX -8% on Tuesday, Mag 7 selling Thursday, OpenAI delay Friday) created persistent pressure on every growth name. Micron's fundamentals have never been stronger. The market's hesitation is about the macro environment in which those fundamentals must translate to sustained multiple expansion.

FedEx's beat-and-sell dynamics deserve separate attention. FDX is the global logistics bellwether for the post-Iran-deal normalization thesis: if Hormuz reopening restores shipping route efficiency, FedEx should benefit from volume recovery on previously disrupted corridors. Q4 results confirm that: international priority package yields rose 16%, FedEx Express revenue hit $21.57 billion. The selloff was entirely about the FedEx Freight spin-off — the transition costs, integration uncertainty, and FY2027 EPS guide width (the range was wide enough to make short-term models imprecise). The underlying operational thesis is intact; the near-term earnings visibility is not.

Carnival's record adjusted net income ($569 million, +20% YoY) at all-time-high customer deposits ($9.0 billion) confirms that the leisure travel consumer is resilient. The sell-the-news reaction is consistent with the pattern from both Q1 2026 (similar beat, similar selloff) and the broader "buy the rumor, sell the news" dynamic that has characterized consumer cyclicals this year. The thesis: Hormuz reopening reduces jet fuel costs, reducing cruise operating costs, which the next quarter's guidance will capture. Carnival's forward margin story is arguably more interesting than the current-quarter report.

Geopolitical Update

Iran: Switzerland Delivers a Roadmap — Now the Details

The week's geopolitical surprise was the positive kind: the Switzerland talks that had appeared to be cancelled (the prior week's search showed the original schedule had been disrupted) actually proceeded on June 21–22 and produced substantive outcomes. According to Al Jazeera and NPR reporting on June 22, the key outcomes were:

1. Roadmap agreement: Both sides agreed on "a roadmap towards reaching a final deal within 60 days." The 60-day clock from June 17 places the hard deadline at approximately August 16, 2026. The Switzerland roadmap adds a process layer: negotiating teams now have an agreed framework for how the remaining issues will be sequenced.

2. IAEA inspections: Iran agreed to invite IAEA inspectors back into the country. This is described by US officials as "a major milestone" and "the first step in permanently denuclearising Iran." The agreement to allow inspections does not resolve what the inspectors will be permitted to see — the intrusiveness question remains disputed.

3. Lebanon deconfliction: A US-Iran agreement to end military operations in Lebanon was included in the Switzerland package, per the CNBC reporting on June 22. This extends the Iran deal's scope beyond the original Hormuz/nuclear framework.

4. Working groups established: Three negotiating tracks created — sanctions, nuclear, and oversight — with teams to begin substantive meetings before the end of June.

Outstanding disputes that remain unresolved: the level of uranium enrichment Iran will be permitted to continue; what happens to Iran's stockpile of highly enriched uranium (20% and 60% enriched material); how intrusive IAEA inspections will be; and the timing and scope of sanctions relief. NPR reported on June 23 that "a US–Iran dispute over nuclear inspections clouds work to finalize a war-ending deal," indicating that even the agreed inspection framework has implementation disagreements.

Net assessment: The Switzerland outcome materially reduces the August 16 deadline risk. A deal that fails because the talks never happened is a different scenario than a deal that fails because the inspection terms prove irreconcilable with both sides having agreed on a roadmap. WTI's continued decline to sub-$71 reflects the market's increasing confidence in the deal-completion scenario (where WTI prices toward $68–74) rather than the deal-collapse scenario (where WTI retests $82–90). The 60-day framework has 50 days remaining as of June 27.

The AI Memory Economy: When the Supplier's Gains Are the Consumer's Losses

The Apple and Microsoft price announcements on Thursday June 25 are the supply-chain confirmation of what Micron's 84.9% gross margin had already quantified. The sequence is worth spelling out precisely:

1. AI data center operators (Microsoft Azure, Google Cloud, Amazon AWS, Anthropic) need HBM3E/HBM4 for AI training and inference at unprecedented scale.
2. HBM production capacity is a multi-year build — TSMC, SK Hynix, Micron, and Samsung cannot add meaningful new capacity in fewer than 18–24 months.
3. The capacity constraint means AI buyers are paying whatever the market will bear, compressing the supply available to non-AI memory buyers.
4. Consumer DRAM and NAND (for MacBooks, iPads, iPhones, Xbox consoles) must be priced from the same constrained pool, at margins reflecting AI-tier pricing.
5. Apple and Microsoft can either absorb the cost (margin compression) or pass it through (price hikes). They chose pass-through.

Microsoft explicitly stated that "memory prices have increased by more than 2.5× and we expect another doubling by the fall of 2027" — a signal that the AI-memory supercycle is not a single-quarter phenomenon but a multi-year structural repricing. The June 2026 consumer electronics price hike cycle is therefore the beginning, not the peak, of the AI memory tax on device buyers. The commodity_supercycle thesis historically applies to extractive commodities (oil, copper, lithium); AI memory is the first digital commodity to exhibit the same inelastic-supply, inelastic-demand dynamics that characterized copper in 2004–2008.

OpenAI IPO Delay: The $1 Trillion Floor and Its Market Implications

The reported delay of OpenAI's IPO until 2027 is a secondary IPO market signal and a primary AI sentiment signal. The mechanics:

OpenAI's advisory team presented management with two paths: (a) list in 2026 at a valuation below $1 trillion (reflecting current private-market comps and SpaceX's post-ATH retreat to $153), or (b) wait until 2027 at a $1 trillion valuation when the AI narrative may have restabilized. Sam Altman reportedly made clear that any proposal below $1 trillion is unacceptable. The decision to wait is rational given that: (1) SpaceX, the closest comparable in terms of market scale, has lost 32% from its ATH in 8 trading sessions (June 16–26, excluding Juneteenth June 19); (2) a $1 trillion IPO during a Nasdaq -4.6% week would price at a discount and leave institutional buyers underwater on Day 1; (3) OpenAI's revenue trajectory (which already crossed $10 billion annualized in mid-2025; by 2026 the run rate exceeds $25 billion) supports a patient 2027 timeline.

The SoftBank impact (-12%) reflects the Vision Fund's locked-in capital: every month OpenAI delays its IPO, SoftBank's Vision Fund 2 (which holds the OpenAI stake) cannot realize its paper gain, reducing SoftBank's ability to fund new investments and pay dividends. Masayoshi Son had reportedly been counting on OpenAI's 2026 listing to trigger a capital recycling event for Vision Fund 2.

The broader implication: the AI IPO window has narrowed. SpaceX's experience — the world's most anticipated technology listing, hitting $225 on Day 1's intraday high and retreating to $153 — suggests that AI valuations at listing are extremely sensitive to the macro backdrop at the time of pricing. If Micron's 84.9% gross margin and $50B Q4 guide can't sustain a Nasdaq rally for a week, the conditions required for a $1 trillion OpenAI listing are more fragile than they appeared in May.

Strategy Scorecard

Winners

Mixed

Losers

MVP of the Week

defensive_rotation — and the margin of victory is the 5.2 percentage point spread between the Dow (+0.60%) and the Nasdaq (-4.60%). In a week defined by three distinct selloff triggers (Tuesday KOSPI crash, Thursday Apple/Microsoft sticker shock, Friday OpenAI IPO delay), defensive sectors provided portfolio stability while growth names absorbed the losses. Utilities, industrials (Caterpillar, Honeywell), healthcare (UnitedHealth, J&J), and consumer staples held or gained while the Nasdaq's heaviest weights — Apple, Microsoft, Alphabet, Meta — fell on AI cost inflation, hot PCE, and AI IPO uncertainty. The 5.2 pp weekly spread is the largest of 2026 and confirms what the strategy's framework has been positioned for since the June 17 FOMC meeting: Kevin Warsh's dropped-forward-guidance and hawkish dot plot have created a rate-uncertainty premium that depresses the long-duration multiple (tech/growth) while leaving the shorter-duration, cash-flow-heavy defensives unaffected. The Dow holding positive in a week where the Nasdaq fell more than 4.5% is the strategy's clearest weekly confirmation of the year.

Next Week Preview: June 29 – July 3, 2026

Economic Calendar

Earnings Reports (Key)

Political / Central Bank

Geopolitical Watchlist

• Iran nuclear inspection terms: The IAEA inspections agreement is the Switzerland meeting's most significant deliverable and its most contested implementation challenge. The US wants inspectors to have access to Iran's entire nuclear infrastructure including military sites; Iran has historically refused access to sites it classifies as sensitive. The specific terms of the inspection protocol — agreed in principle in Switzerland — need to be finalized before the 60-day deadline. Any breakdown in inspection term negotiations will be reported first in IAEA press releases and will move WTI within hours.
• Uranium enrichment question: Iran's position on whether it will be permitted to continue enriching uranium (Iran had been enriching to 20% before the original JCPOA; the JCPOA itself limited Iran to 3.67% enrichment) or must halt enrichment entirely under the new deal is the central unresolved issue. The US position (zero enrichment) and Iran's position (limited enrichment as a sovereign right) represent the hardest negotiating gap remaining. A compromise at 5% enrichment with robust IAEA oversight would be a deal-sealing development.
• SPCX inclusion dynamics (July 7): At $153 per share versus the $225 ATH at which most Nasdaq-100 inclusion analyses were conducted, the forced passive-buying event is smaller in dollar terms but still the largest forced-buy quantum for any single security in Nasdaq-100 history at these absolute price levels. Watch for institutional pre-positioning in the June 30 – July 6 window: the last three days before inclusion are historically associated with pre-positioning that can lift the inclusion-day price.
• Gold at $4,000 support test: Gold's intraday break below $4,000 on Thursday June 25 (the first since November 2025) was partially recovered by Friday's close (~$4,080). The $4,000 level needs to hold as support on Monday June 29 or the gold bull thesis faces a sustained test; the next technical support is approximately $3,900–$3,950 (October 2025 levels; the September 2025 highs were closer to $3,820–$3,870).

Monday Setup

Scenario A: NFP below 120K + Iran working groups report progress (~25% probability)

Non-farm payrolls print below 120,000 (or unemployment rises to 4.5%), removing the September rate-hike catalyst; simultaneously, Iran and US negotiating teams report productive working-group meetings before the week's close. WTI tests $66–70 on deal-momentum acceleration; the 10-year yield falls toward 4.35–4.40%, relieving rate pressure on growth multiples; the Nasdaq recovery erases part of this week's -4.60%; SPCX stabilizes at $155–165 ahead of the July 7 inclusion. Gold rebounds toward $4,100–$4,150 as the rate-hike tail risk moderates. Action: release defensive sleeve in momentum_crash_hedge; restore ai_infrastructure_layer to 85%; add semiconductor_value on any Micron weakness; reduce defensive_rotation toward neutral as growth/defense spread narrows; reduce cash to 10%.

Scenario B: NFP in-line (145–175K) + Iran progress neutral (~50% probability)

The most likely scenario: NFP approximately in consensus range keeps September hike probability alive but not dominant; Iran working groups begin without dramatic announcements. WTI oscillates $69–74; 10-year holds 4.50–4.55%; Nasdaq stabilizes after this week's -4.60% selloff (modest technical bounce off oversold conditions); S&P holds 7,300–7,450 range; SPCX begins pre-inclusion positioning toward $160–170. Action: hold defensive_rotation at full weight; hold ai_mega_ecosystem at 60%; maintain warflation_hedge at 8%; begin rebuilding semiconductor_value position if Micron tests $1,100; hold cash at 13–15%.

Scenario C: NFP above 200K + wages hot + PCE follow-through (~25% probability)

Strong jobs data makes September rate hike the base case (>60% probability); the market is simultaneously absorbing hot June PCE in this scenario, as July's reading would also reflect some residual energy-from-May. WTI bounces toward $75–79 if Iran talks show no progress; 10-year tests 4.65–4.75%; Nasdaq tests 24,500–25,000; S&P tests 7,200–7,300; gold rebounds from below $4,000 on stagflation-like readings (hot wages + decelerating growth); SPCX inclusion runs into a risk-off market and struggles at $145–155. Action: activate crisis_rotation; restore gold_bug to 60% (stagflation setup where gold can outperform despite higher real rates if growth slows); hold defensive_rotation at full weight; reduce ai_mega_ecosystem to 40%; add treasury_safe short-duration positions; raise cash to 18–20%.

Position Sizing

• Defensives (defensive_rotation): hold at full weight through NFP; the Dow/Nasdaq spread trade is the defining market structure of June 2026
• Japan (japanese_sogo_shosha, japan_industrial_finance): hold at full weight; BoJ normalization + Hormuz deal + TSE governance is a three-year thesis, not a one-week thesis
• Memory/Semiconductors (semiconductor_value, infrastructure_reshoring): hold; Micron's $50B Q4 guide is the floor under the thesis regardless of weekly price action
• SpaceX (pre_ipo_innovation_funds): hold at 40% reduced weight through July 7 inclusion; the forced-buy event remains but the magnitude is reduced from original estimates
• Energy (warflation_hedge): 8% tail hedge only; the base case is deal completion and WTI toward $68–74
• Gold (gold_bug): 20% weight; hold for technical recovery attempt; below $3,950 on any session warrants further reduction
• AI broadly (ai_mega_ecosystem): 60% weight; the OpenAI delay and Mag 7 underperformance suggest the AI narrative needs a new catalyst; wait for NFP direction before rebuilding
• Cash: 13–15%; prioritize NFP clarity on Thursday July 3 before deploying

---

The PCE Time Machine: Why Today's Inflation Data Is Yesterday's News — and Tomorrow's Is Already Written

May PCE came in at 4.1% headline and 3.4% core on Thursday, the highest readings since April 2023 and October 2023 respectively. The bond market barely moved. The S&P 500 closed down 0.01%. Here is why the most important inflation data release of the month produced almost no reaction — and what the actual inflation trajectory of 2026 is likely to look like.

May PCE is a photograph of the Iran war. The Iran war ended June 17.

The Personal Consumption Expenditures price index measures price changes in the goods and services Americans buy in a given month. The May data released June 25 measures prices as of approximately May 31, 2026. That is a world where:

• The Strait of Hormuz had been partially or fully disrupted since late January 2026
• WTI peaked near $106/barrel in early May 2026 and averaged approximately $100–102/barrel for the month (the peak conflict period)
• LNG prices in Asia were elevated, increasing manufacturing input costs globally
• Supply chains for oil-dependent industries (petrochemicals, plastics, shipping) were pricing in weeks-long route detours around the Cape of Good Hope
• The Iran war's economic shock was compounding with the AI-memory chip cost surge already in the pipeline

As of June 26, 2026:
- The Hormuz Strait has been physically open for nine days
- WTI is below $71 — a decline of approximately $17–24 from the May average, depending on which week of May you anchor to
- 54 supertankers that had been queued in the Gulf are now transiting; the shipping backlog is clearing
- LNG route normalization is following oil's disinflationary trajectory

The oil-to-PCE transmission mechanism has a predictable lag.

Based on historical oil-price-to-CPI/PCE pass-through data, the relationship works as follows: every $10/barrel sustained decline in WTI translates to approximately 0.2–0.35 percentage points off headline PCE/CPI within 4–8 weeks, depending on the mechanism (direct gasoline prices respond faster; heating oil, electricity, and manufacturing inputs take 8–12 weeks).

The implication of a $17–24 per barrel decline from the May average:
- Estimated headline PCE contribution from oil normalization: approximately -0.35 to -0.85 percentage points within 2 months
- June PCE (releasing approximately July 30): the June average WTI of approximately $73–76 vs May average of $100–102 suggests energy's contribution to headline PCE will reverse sharply
- July PCE (releasing approximately August 28): if WTI holds at $68–74 (the deal-completion scenario), the July reading could show further disinflation

Bond markets are doing exactly this math. The 10-year Treasury yield fell approximately 8 basis points on the week despite Thursday's hot PCE print, the opposite of what a naive model would predict (PCE at 3.4% core, highest since October 2023, with the Fed having just signaled a hawkish dot plot, would historically move the 10-year up by 15–25 bps). The counter-intuitive move tells you: sophisticated bond buyers are not pricing May's 4.1% PCE as a steady-state. They are pricing the June data they can already estimate from WTI's current level.

But here is the twist the simple narrative misses: the AI memory price shock is building into core PCE just as energy fades from headline.

Core PCE (which excludes food and energy) was 3.4% in May. The components of core PCE include durable goods — computers, electronics, appliances. Apple's MacBook price hikes of $200–$300 (+18–20%) and Microsoft's Xbox increases of $100–$150 (+15–20%) are durable goods. When consumers buy these products, the higher prices are captured in the "durable goods" line of core PCE.

Microsoft explicitly stated: "memory prices have increased by more than 2.5× and we expect another doubling by the fall of 2027." That statement is not a price-guidance forecast — it is a cost-of-goods timeline. The inflation Warsh is fighting from energy (which is transitory, because oil prices reflect a specific geopolitical event that has been resolved) is about to be partially replaced by AI-memory inflation in core goods (which is structural, because HBM production capacity cannot be doubled in 12 months).

The PCE Time Machine therefore runs in two directions simultaneously:

Backward: May PCE at 4.1% is measuring the past — the Iran war's oil price peak. That data is already obsolete. The June and July readings will show dramatic disinflation in the energy component that the May print cannot capture. Markets are not wrong to shrug at 4.1% when WTI is at $71.

Forward: The core PCE in June, July, and August will increasingly reflect AI-memory-driven goods price inflation: higher MacBook, iPad, Xbox prices showing up in the durable goods component. Ironically, as the Iran war's contribution to headline PCE deflates, the AI supercycle's contribution to core PCE inflates. Warsh's 3.6% headline PCE forecast for 2026 may be too high for energy but too low for core goods.

The historical parallel that fits this dynamic best is 2005–2006. After Hurricane Katrina drove oil to $70+ (then a shock level) in late 2005, gasoline prices contributed approximately 1.0–1.5 percentage points to headline CPI/PCE. As oil prices normalized in 2006, headline inflation fell — but core PCE simultaneously rose as housing costs (driven by the Fed's low-rate-induced housing boom) replaced energy as the dominant inflation vector. The Fed, targeting headline PCE and then core, missed that the inflation source had rotated. The result was a persistent core PCE above 2% even as oil-driven headline softened, and a Fed that tightened too little in retrospect because the headline PCE trajectory looked encouraging.

In 2026, the rotation is from oil-war energy inflation to AI-memory-cost goods inflation. The Fed is still fighting the energy version. The AI version is already priced into Apple's MacBook shipping today.

What investors should track:
1. June CPI (releasing approximately July 14): the energy component's collapse will be visible in the headline; the core reading (especially durable goods) will give the first read on AI-memory price pass-through
2. June PCE (releasing approximately July 30): if core PCE rises despite the Hormuz energy normalization, it signals that the AI goods inflation is already dominating the energy disinflation
3. Micron's Q4 FY2026 (reporting approximately September 2026): the Q4 guidance of $50 billion in revenue at 86% gross margins — if confirmed — will extend the memory price supercycle through at least December 2026 and into Q1 2027, meaning the AI goods inflation vector doesn't peak until Micron's margins peak

The commodity_supercycle thesis has historically been applied to extractive commodities — copper in 2004–2008, oil in 2007–2008, lithium in 2021–2023. AI HBM memory in 2025–2027 is exhibiting every characteristic of a supercycle: inelastic supply (HBM fabs cannot be built in fewer than 18–24 months), inelastic demand (AI builders need HBM or cannot train their models), and pricing dynamics that allow the supplier (Micron) to charge what the market will bear (84.9% margins). The PCE reading of that supercycle will look different from the oil-supercycle reading because HBM shows up in core goods, not in the energy component — which means it is structurally more persistent and more difficult for a central bank targeting 2% core to counteract without destroying the AI buildout that is actually driving productivity growth.

The PCE time machine's most unsettling implication: Warsh's 3.6% headline PCE forecast for 2026 may be directionally right for the wrong reasons — not because the Iran war's oil shock persists, but because the AI memory shock is arriving just as energy inflection deflates.

---

Sources:
- Stock Market Today June 23, 2026 — TheStreet
- Nasdaq Composite Posts Fifth Losing Session Friday as Chip Stocks Tumble — CNBC
- Stock Market Today June 25, 2026 — TheStreet
- Stock Market Today June 26, 2026 — TheStreet
- Kospi Index Slides 4.6% With Samsung, SK Hynix Falling on Chip Concerns — Bloomberg
- Wall Street Gets Trampled by AI Sell-Off; South Korean Market Plunges 10% — CNN Business
- Kospi Crashes Nearly 10% on Massive Tech Sell-Offs — The Korea Herald
- AI Stock Selloff Hits Nasdaq 2.2%: KOSPI Circuit Breaker Signals a Divided Market — TechTimes
- Earnings Call Transcript: Micron Tops Q3 2026 Estimates, Shares Jump 14.6% — Investing.com
- Micron Q3 Revenue Beats Forecasts at $41.46 Billion — BreakingTheNews
- Micron (MU) Earnings Report Q3 2026 — CNBC
- Micron Technology Fiscal Q3 2026 Earnings Call Prepared Remarks — Micron Investor Relations
- Micron Stock Hits $1,213 Before Q3 Earnings — Vantage Markets
- FedEx Q4 FY26 Slides: Beats Estimates But Shares Fall on Outlook — Investing.com
- FedEx (FDX) Q4 2026 Earnings — CNBC
- Carnival Corporation Releases Q2 2026 Financial Results — Alphastreet
- Darden Restaurants Q4 2026 Earnings — CNBC
- PCE Inflation Report May 2026: Core Inflation Rate Hit 3.4%, Highest Since October 2023 — CNBC
- The Fed's Preferred Inflation Gauge Shows Prices Rising at Fastest Pace in 3 Years — CBS News
- Apple, Microsoft Hike Prices Over Surging Chip Costs — Al Jazeera
- The AI Price Shock Is Here: Apple and Microsoft Hike Prices — Axios
- Apple and Microsoft Hike Prices as AI Crunches Global Memory Chip Supply — CBC News
- SpaceX Will Join Nasdaq-100 — CNBC
- SpaceX Will Join the Nasdaq-100 Index on July 7, 2026 — WEEX
- OpenAI Is Reportedly Delaying Its IPO: Here's When Kalshi Traders Think It Will Announce — CNBC
- AI Trade Hits a Wall Amid Report That OpenAI Will Delay IPO Until 2027 — Yahoo Finance
- SoftBank Shares Tumble After Report of OpenAI's IPO Delay — Bloomberg
- OpenAI IPO Delay Sends SoftBank Shares Plunging Over 12% — TradingKey
- U.S., Iran Agree on Roadmap for Final Deal and Plan to End Military Operations in Lebanon — CNBC
- Key Outcomes of Iran-US Talks in Switzerland — Al Jazeera
- The U.S. and Iran Agree to a 'Road Map' for a Final Deal, Mediators Say — NPR
- A U.S.-Iran Dispute Over Nuclear Inspections Clouds Work to Finalize a War-Ending Deal — NPR
- What the U.S. and Iran Agreed and Disagreed on First Day of Talks — Al Jazeera
- Gold Prices Today, Thursday June 25, 2026: Prices Fall Below $4,000 for the First Time Since Nov. '25 — Yahoo Finance
- Nikkei 225 Falls -2.8% to Below 70,400 as Global Tech Sell-Off Reverses Thursday's Historic Gains — BBN Times
- Nikkei 225 Today: Index at 69,400 After Chip-Led Selloff — Vantage Markets
- SpaceX (SPCX) Set for Nasdaq Inclusion Ahead of July 2026 — GuruFocus

---

Disclaimer

This report is produced for informational and educational purposes only and does not constitute investment advice, a solicitation, or a recommendation to buy or sell any security. All data cited reflects information available as of the publication time noted above. Market conditions and geopolitical developments may change materially before or during the trading session. Futures and pre-market levels are indicative only and are not guaranteed opening prices. Past performance of any strategy referenced is not indicative of future results. Consult a qualified financial advisor before making investment decisions.