The S&P 500 closed the week +1.24%, its gain concentrated in AI-infrastructure names even as the Dow finished red (-0.50%) on a mid-week Iran-driven selloff it never fully recovered from. The dominant driver was conviction in the AI-infrastructure buildout — Broadcom’s expanded Apple chip deal, Meta’s cloud pivot, and SK Hynix’s record Nasdaq debut — proving strong enough to absorb both a genuine US-Iran ceasefire collapse and a meaningfully more hawkish Fed. That resilience came with a real policy cost: 2026 rate-hike odds jumped to 59% by week’s end, setting up Tuesday’s CPI print and Chair Warsh’s congressional debut as the decisive next tests.

1. AI-Semiconductor Conviction Wins the Week: Broadcom-Apple Deal, Tuesday’s Valuation Scare, and SK Hynix’s Record $26.5B Nasdaq Debut

The core facts:Monday opened with Broadcom disclosing a multi-year custom-chip extension with Apple through 2031; Tuesday reversed hard as Samsung’s beat-but-sold-off earnings and a Reuters report that China’s DeepSeek is designing its own inference chip erased over $100B of Samsung value and dragged Intel (-9.66%), Marvell, SanDisk, KLA, and Lam Research down 6-11%. Wednesday, Apple and Broadcom finalized an expanded $30B+ agreement (15B+ US-made chips through 2031) and reports of Chinese H200 approval lifted Nvidia even as the broader tape sold off on Iran. Thursday the SOX index jumped 3.1% ahead of SK Hynix’s Friday listing; Friday SK Hynix priced at $149, opened at $170 (+14%), and closed with a $1.25 trillion market cap on 7x order-book oversubscription — the largest-ever US listing by a foreign company. The Trump administration separately eased UAE chip-export rules Friday, opening license-free Nvidia/AMD sales to Gulf AI buyers.

Why it matters:The week is a live stress-test of whether capital markets will keep financing the AI buildout at current valuations, and it answered emphatically yes — but only after a genuine mid-week scare exposed how thin the AI-equipment/memory complex’s margin for disappointment has become. Technology closed the week +2.81% (see sector rotation table in Section B), a number that flattens a week that actually swung from a >$100B single-day Samsung wipeout to a $1.25T oversubscribed debut. The UAE export easing is a structural, not one-day, tailwind that compounds the demand-durability signal.

What to watch:SK Hynix’s first session under its permanent SKHY ticker Monday; Intel’s July 23 earnings for foundry-yield and data-center-share commentary; hyperscaler capex guidance in late-July earnings for confirmation the demand signal holds.

↑ back to summary

2. US-Iran Ceasefire Collapses, Then Markets Shrug It Off: From Tanker Strikes to a Fragile De-Escalation

The core facts:Iran struck at least two tankers near the Strait of Hormuz Tuesday, prompting the US Treasury to revoke Iran’s crude-export license; WTI jumped 5.32% that session. Wednesday, US Central Command struck roughly 90 Iranian targets, President Trump declared the ceasefire memorandum “over,” and oil surged a further 5.9-6.9% as nine of eleven S&P sectors sold off (Dow -1.09%). Thursday brought a second day of US strikes and Iranian retaliation against Kuwait/Bahrain targets, yet WTI fell 2.3% and equities rallied (S&P +0.81%) as markets priced a “bumpy but real” path to de-escalation. By Friday, oil had eased further (WTI -0.93%) on continuing technical talks, even as the IEA flagged a 1.2M bpd 2026 demand contraction.

Why it matters:The gap between Wednesday’s genuine risk-off (nine-of-eleven-sector selloff, gold failing to catch a haven bid) and Thursday-Friday’s shrug is itself the signal: markets are treating the escalation as contained and reversible rather than a durable supply shock, a read that remains fragile given Hormuz traffic (roughly 20% of global oil volume) stays well below pre-conflict levels. Energy still finished the week as the leading S&P sector (+3.50% — see Section B), even as oil gave back most of its mid-week spike.

What to watch:Whether Hormuz tanker traffic is physically disrupted again, and progress in ongoing US-Iran technical talks; any renewed Iranian threat against export infrastructure would force a rapid repricing of this week’s calm.

↑ back to summary

3. Fed’s Hawkish Drift Under New Chair Warsh: FOMC Minutes, Williams’ AI-Inflation Warning Push 2026 Hike Odds to 59%

The core facts:Monday, Fed Governor Waller called forward guidance “more art than science,” visibly in tension with Chair Kevin Warsh’s data-only posture. Wednesday’s June FOMC minutes — Warsh’s first meeting — showed 9 of 18 officials eyeing a hike, with staff flagging AI-infrastructure demand as a structural inflation input. Thursday, NY Fed President Williams, traditionally a centrist, named AI-driven demand his “top inflation risk” and said the Fed “won’t look through” it; Polymarket’s 2026 hike-odds jumped 11 points in a single session to 59%. Friday, the Fed’s first Monetary Policy Report under Warsh vowed an “absolute commitment” to 2% inflation, setting up his House and Senate testimony next Tuesday and Wednesday. Warsh also named task-force leaders including Marc Andreessen (labor/AI) and Doug McMillon (data quality) Thursday.

Why it matters:A centrist voter (Williams) explicitly validating the hawkish minority meaningfully raises the odds Warsh’s committee tightens further this year, and it directly entangles Fed policy with the same AI-capex cycle powering the week’s equity rally (see Story 1) — an AI slowdown would now hit both growth and the policy path simultaneously. Yet a 30-year Treasury auction Thursday cleared at the richest yield since 2007 with blowout foreign demand (77.7% indirect bidders), suggesting credit markets aren’t yet panicking about the hawkish repricing.

What to watch:Tuesday’s June CPI print and Warsh’s same-day House testimony (Senate follows Wednesday) are the two highest-probability catalysts for the next repricing of 2026 hike odds in either direction.

↑ back to summary

4. Meta’s AI-Cloud Pivot: “Meta Compute,” In-House Iris Chip, and a $9.2B Canadian Data Center Anchor the Stock’s Best Week Since 2024

The core facts:Wednesday, Meta committed C$13B (~$9.2B) to a 1-gigawatt Alberta AI data center — its first in Canada. Friday, shares rose 5.97% to $669.21, extending the week’s gain to roughly 15%, on continued enthusiasm for a planned “Meta Compute” AI cloud business that would compete directly with AWS, Azure, and Google Cloud, plus disclosure of an in-house “Iris” AI chip targeting 14GW of capacity at roughly half prior Street cost assumptions. Meta ended the week as the second-biggest weekly gainer among mega-caps (see weekly movers table in Section B).

Why it matters:Meta’s disclosure that its in-house chip costs run roughly half of Street assumptions is a genuine capex-efficiency signal that, if it holds, eases the exact investor concern — unmonetized, escalating AI capex — that has periodically hit the stock all year. It also demonstrates hyperscaler AI-infrastructure spending is still accelerating even as the broader software complex de-rates (see Story 5), reinforcing that investors are discriminating within the AI trade, not abandoning it.

What to watch:Confirmation of the Iris chip’s September production start; Meta’s Q2 earnings call for formal capex guidance and Meta Compute’s monetization timeline.

↑ back to summary

5. Valuation Reset Sweeps High-Multiple Software: Palo Alto, Palantir, Dell, Marvell, and Netflix All Slide Even as AI Hardware Rallies

The core facts:Palo Alto Networks hit an all-time high Monday on a string of analyst upgrades, then fell 5.5% Wednesday and another 3.67% Friday on profit-taking despite Evercore trimming (not abandoning) its target; the stock finished the week -6.36%. Palantir fell 4% Wednesday, extending a 29-40% pullback from its 2026 highs. Dell and Marvell each fell 3-4% Friday on AI-server margin and custom-ASIC competition concerns. Netflix fell 2.78% Friday to a fresh multi-month low, down 5.51% on the week, on subscriber-engagement worries and reports it may launch live TV channels; Bernstein cut its target to $100 from $110.

Why it matters:The divergence between AI-infrastructure hardware (rallying, see Story 1) and high-multiple software/platform names (de-rating) shows investors discriminating sharply within the broader technology trade rather than treating it as one basket — a rotation likely to persist as long as rate-cut conviction stays low (see Story 3). PANW and Dell both appear among the week’s top mega-cap decliners (see Section B), underscoring the pattern isn’t isolated to one name.

What to watch:Whether the valuation-reset selloff broadens further into cybersecurity and enterprise software; Netflix’s July 16 earnings for subscriber and engagement metrics.

↑ back to summary

6. SpaceX’s Nasdaq-100 Debut Reveals Credit Stress as Bonds Trade at Junk-Equivalent Spreads

The core facts:SpaceX joined the Nasdaq-100 Tuesday under a new fast-track rule just 15 trading days after its June 12 IPO, but shares fell 6.34% that day as most of an estimated $4.3B in forced passive buying had already been absorbed — a classic “buy the rumor, sell the news” pattern. The stock kept sliding into Friday (-4.51%), down 10.31% on the week (the week’s worst mega-cap decliner — see Section B), as nominally BBB-rated SpaceX bonds traded at spreads consistent with BB (junk) credit, tied to $17.5B in debt assumed from its ~$60B xAI/X acquisition and a reported $30B annualized AI-unit cash-burn rate.

Why it matters:A roughly $1.9 trillion-valued company’s debt trading at junk-equivalent spreads is a genuine credit-stress signal, not noise. Because SpaceX’s financing has become a bellwether for how capital markets price the broader AI-infrastructure buildout (alongside Story 1), further deterioration here would raise real questions about financing conditions for other capital-intensive AI bets.

What to watch:Any formal credit-rating action from Moody’s, S&P, or Fitch; whether SpaceX’s bond spreads stabilize or widen further.

↑ back to summary

7. Vertex Pharmaceuticals to Acquire Crinetics for $10 Billion, One of 2026’s Largest Biotech Deals

The core facts:Vertex agreed Monday to acquire Crinetics Pharmaceuticals for $85.00/share in cash (~$10B, ~$8.8B net of Crinetics’ cash), unanimously approved by both boards and expected to close Q3 2026. Crinetics markets PALSONIFY, the first once-daily oral acromegaly therapy; Vertex projects combined peak revenue above $5B, becoming accretive to non-GAAP operating income in 2029, financed via a $4.5B Bank of America/Morgan Stanley bridge facility. Crinetics shares surged over 100%; Vertex traded roughly 0.5% lower.

Why it matters:This is one of the largest biotech acquisitions of 2026 and signals continued large-cap pharma appetite for commercial-stage rare-disease assets over organic R&D. Vertex’s modest decline reflects standard deal-financing digestion (debt-funded, multi-year accretion) rather than strategic disapproval; Crinetics’ full premium pop confirms an uncontested bid.

What to watch:Regulatory clearance and the Crinetics shareholder vote ahead of the targeted Q3 2026 close.

↑ back to summary

8. Trump Rings NYSE/Nasdaq Bell From the Oval Office, Launches “Trump Accounts” With a $6B Dell Family Pledge

The core facts:President Trump rang the opening bell of both the NYSE and Nasdaq from the Oval Office Monday — the first time either bell has been rung from the White House — marking the launch of “Trump Accounts,” a new tax-advantaged investment program for American children. Michael and Susan Dell attended and pledged more than $6 billion to the program; Trump promoted Dell computers directly during the ceremony, and Dell shares rallied 4.43% that day (extending to +10.31% on the week — see Section B).

Why it matters:Beyond Dell’s single-stock pop, a nationwide tax-advantaged child investment program is a material new federal financial-policy initiative with potential long-run implications for retail investment flows and household balance-sheet formation, while the White House staging itself signals this administration’s active use of capital markets as a policy and messaging tool.

What to watch:Enrollment and funding details as Trump Accounts roll out; whether other companies pursue similar high-profile funding pledges.

↑ back to summary

9. State AGs Prepare Antitrust Suit to Block $110B Paramount Skydance-Warner Bros. Discovery Merger

The core facts:Reuters reported Thursday that California, New York, Pennsylvania, and roughly ten other states are finalizing an antitrust lawsuit to block Paramount Skydance’s approximately $110 billion acquisition of Warner Bros. Discovery, arguing the deal would harm entertainment and news competition. Paramount Skydance shares fell 6% on the report; the company has publicly denied the allegations and is separately seeking dismissal of earlier private litigation.

Why it matters:This would be one of the largest media mergers on record, and a coordinated multi-state suit layered on top of the DOJ’s existing Second Request review materially raises the odds of a prolonged legal fight or forced deal restructuring — a reminder antitrust enforcement remains an active constraint on large-scale consolidation even in a generally permissive M&A environment.

What to watch:Formal filing of the state lawsuit, expected within weeks; any parallel DOJ Second Request outcome or divestiture requirement.

↑ back to summary

10. Apple Sues OpenAI Alleging Trade-Secret Theft Tied to AI Hardware Ambitions

The core facts:Apple filed suit Friday in Northern California federal court against OpenAI, its io Products hardware unit, and two former Apple employees, alleging OpenAI leadership directed a pattern of trade-secret theft to accelerate its consumer AI hardware ambitions — including use of Apple project code names in recruiting and retention of confidential product documents. Apple’s complaint states more than 400 former Apple employees now work at OpenAI, which acquired Jony Ive’s io Products for roughly $6.5 billion.

Why it matters:The suit is a direct escalation in the Apple-OpenAI consumer-AI-hardware rivalry and signals how seriously incumbent hardware makers are treating personnel and IP leakage to AI-native competitors — a dynamic relevant to any public company competing for AI talent, even though OpenAI itself has no direct public-market read-through.

What to watch:OpenAI’s formal response and whether Apple seeks a preliminary injunction that could delay or constrain OpenAI’s hardware product timeline.

↑ back to summary

ISM Services PMI Eases to 54.0 in June as Hiring Jumps and Price Pressures Cool (ISM, Mon, Jul 6)

What they’re saying:The ISM Services PMI eased to 54.0 in June from 54.5, matching consensus. Business Activity slowed to 55.4 from 57.7 and New Orders eased to 55.1 from 57.3. The Employment Index jumped to 51.2 from 47.9 — its largest increase since 2024 and the first expansion in four months — while the Prices Index eased to 67.7 from 71.3, a four-month low.

The context:The report cuts both ways for the week’s macro debate: decelerating services growth alongside the sector’s first hiring expansion in months complicates any clean “labor is cooling” read, while the cooler prices component offers the Fed modest reassurance even as Thursday’s Williams remarks and Friday’s Monetary Policy Report kept the tone hawkish (see Section C).

What to watch:July ISM Services PMI due Aug. 5, for whether the hiring rebound holds against this week’s other mixed labor signals.

Initial Jobless Claims Fall to 215K, Beating Estimates as the Labor Market Holds Resilient (Dept. of Labor, Thu, Jul 9)

What they’re saying:Initial jobless claims for the week ended July 4 came in at 215,000, below the 218,000 consensus and down from the prior week’s revised 217,000. The four-week moving average eased to 218,750 from 222,500, continuing a gradual downtrend.

The context:The beat reinforces labor-market resilience even as pockets of softness show elsewhere (ADP’s weekly pulse, participation concerns below); landing the same day as Williams’ hawkish AI-inflation remarks, it added to the case the Fed sees no urgency to ease, reinforcing the week’s hawkish repricing (see the macro synthesis above).

What to watch:Continuing claims (1.814M, still cycle-elevated) for signs re-employment is slowing; the July nonfarm payrolls report due in early August.

Existing Home Sales Fall to 4.09M in June, Missing Estimates as Median Price Hits Record $440,600 (NAR, Thu, Jul 9)

What they’re saying:Existing home sales fell 2.4% month-over-month to a seasonally adjusted annual rate of 4.09 million units, below the 4.20 million consensus and down from May’s 4.19 million pace. The median existing-home price climbed to a record $440,600, up 1.8% year-over-year.

The context:NAR’s own affordability index actually improved to 102.3 from 95.5 a year ago and sales remain up 2.8% year-over-year — a two-speed housing market of improving affordability but rate-sensitive month-to-month demand. Realtor.com separately trimmed its 2026 sales and price-growth forecasts this week, citing Middle East-driven rate pressure (see Iran arc in Section C).

What to watch:New and pending home sales over the next two weeks; the 30-year mortgage rate (6.49% as of Thursday) for any relief that could unlock demand.

Pantheon Warns of Unemployment “Snapback” Risk as Labor-Force Participation Hits a 50-Year Low (Pantheon Macroeconomics, Wed, Jul 8)

What they’re saying:Pantheon Macroeconomics flagged that the labor-force participation rate’s slide to 61.5% — the lowest outside the pandemic since 1976, after a 720,000 drop in the labor force in June concentrated among prime-age workers — leaves the job market vulnerable to a “snapback” in unemployment in the second half of 2026 if discouraged workers re-enter.

The context:A falling participation rate has been artificially flattering the headline unemployment rate; any reversal would mechanically push unemployment higher even without net job losses, complicating the Fed’s read on labor-market slack precisely as hike odds are rising (see the macro synthesis above and the Fed thread in Section C).

What to watch:July’s employment report (Aug. 1) for early signs of a participation rebound.

May Trade Deficit Widens to $77.6B, Biggest Gap Since March 2025, as Import Surge Weighs on Q2 GDP (BEA, Tue, Jul 7)

What they’re saying:The trade deficit widened to $77.6 billion from a revised $54.6 billion in April, roughly in line with the $78.5 billion estimate. Imports climbed 3.3% to a more-than-one-year high of $395.3 billion — led by pharmaceuticals, cell phones, crude oil, and passenger cars — while exports fell 3.2% to $317.7 billion.

The context:A widening deficit subtracts directly from GDP’s net-exports component; the Atlanta Fed’s GDPNow eased to 1.3% from 1.4% partly reflecting the drag, even as the deficit remains 40.6% narrower than the same period in 2025 — tariff-driven import front-running remains the dominant, ongoing distortion rather than a new structural shift.

What to watch:The BEA’s advance Q2 GDP estimate (late July) for confirmation of the net-export drag; June trade data due mid-August.

1. PepsiCo (PEP): -3.26% | Beats on EPS and Revenue, But Margin Contraction and a Cautious North America Outlook Weigh

The Numbers:Revenue of $24.18B beat the $23.95B estimate (+0.98% surprise); adjusted EPS of $2.20 beat the $2.19 estimate; GAAP EPS of $2.18 beat the $2.16 estimate. Released BMO, Thursday, Jul 9. The company reaffirmed full-year 2026 guidance of 2-4% organic revenue growth and 4-6% core constant-currency EPS growth.

The Problem/Win:Despite the top- and bottom-line beat, core operating margins contracted roughly 40 basis points and North American demand stayed weak as inflation-pressured consumers pulled back, offsetting stronger international volume growth; investors focused on the margin pressure and cautious tone rather than the headline beat.

The Ripple:The reaction extends a pattern visible across the week’s consumer names — Costco (-4.24% Thursday on a comp-sales deceleration) and Netflix (-2.78% Friday on engagement concerns, see Section C) were both marked down on demand/margin softness even where headline metrics cleared the bar.

What It Means:Reaffirmed guidance suggests management doesn’t see structural deterioration, but the market is signaling reduced tolerance for margin softness in staples names amid a still-pressured North American consumer.

What to watch:Commentary on North American pricing and volume trends at the next earnings call; whether input-cost inflation pressures margins further in H2.

Q1 2026 earnings season is complete. Q2 2026 earnings season opens Tuesday, July 14, with the four largest U.S. banks reporting before the bell — the traditional kickoff that sets the tone for the broader season.

JPMorgan Chase (JPM) — BMO, Tue Jul 14 — net interest income guidance, trading and investment-banking revenue trends, and credit-provision levels across consumer and commercial books will set the tone for the sector.

Bank of America (BAC) — BMO, Tue Jul 14 — net interest income trajectory and consumer-banking margin trends as rate-cut expectations continue to shift.

Goldman Sachs (GS) — BMO, Tue Jul 14 — whether the firm’s investment-banking and M&A momentum (global deal-making tracked roughly $1.2 trillion in Q2) extends into results, alongside Global Banking & Markets and Asset & Wealth Management growth; the firm’s board is also expected to act on a previously announced quarterly dividend increase to $5.00 from $4.50 per share.

Wells Fargo (WFC) — BMO, Tue Jul 14 — net interest income guidance and expense discipline following the lifting of its asset cap.

Citigroup (C) — BMO, Tue Jul 14 — progress on its multi-year restructuring alongside capital-markets and international consumer-banking trends.

The money-center banks’ results will set the tone for broader S&P 500 reporting, which accelerates through late July.

1. Does Tuesday’s CPI validate or reverse the week’s hawkish repricing? After hike odds jumped to 59% on Fed rhetoric alone, a cool print would force a rapid unwind heading into the July 28-29 FOMC, while a hot one would cement it.

2. Can Chair Warsh’s testimony reconcile the Fed’s still-unsettled communication framework? Waller’s “more art than science” stance, Williams’ AI-inflation warning, and Warsh’s own data-only posture produced three different signals in one week; his House and Senate testimony is the first chance to unify them.

3. Does the AI-infrastructure conviction trade survive contact with money-center bank earnings? JPMorgan, Bank of America, Goldman Sachs, Wells Fargo, and Citigroup kick off Q2 season Tuesday; trading-revenue commentary tied to this week’s volatility (Iran, SK Hynix’s debut) is an early test of whether the rally’s breadth extends into financials.

4. Does the SpaceX credit-stress signal spread to other AI-capex-heavy balance sheets? With SpaceX bonds trading at junk-equivalent spreads, similar strain surfacing elsewhere would validate the tension this week’s Chart of the Week (below) flagged between AI equity euphoria and credit-market skepticism.

WHY THIS CHARTThis chart’s warning that AI credit markets are pricing an exit even as equities price the upside proved directly prescient this week — SpaceX’s bonds traded at junk-equivalent spreads by Friday (see Section C, Story 6), while the same AI-infrastructure trade delivered SK Hynix’s record Nasdaq debut and Broadcom’s expanded Apple chip deal (Story 1). No other candidate this week captured both halves of that central tension as cleanly.

ORIGINAL CHART ANALYSIS — FROM MONDAY’S MIBEquity markets and credit markets are reading the same AI balance sheets and reaching opposite conclusions — and history says the bond desk is the one to believe. Since January 2025, five-year CDS on investment-grade AI names has decoupled from the broad CDX IG BBB index, swinging to roughly +5bp and widening while the benchmark grinds ~15bp tighter — a 20bp gap that opened even as equity indices printed records (Panel B). Two structural facts argue it has further to run. First, the leverage is new and unrepriced: after a net ~$40bn paydown in 2024, AI-linked issuers levered up by a record ~$340bn in 2025, roughly double the 2020 peak (Panel A). Second, the plumbing is reflexive: circular financing accounts for ~58% of lab and ~57% of hyperscaler commitments — and an extraordinary ~96% of chip-maker commitments, the most-exposed link, whose forward book is almost entirely equity-for-purchase deals struck with their own customers. Their demand has no independent buyer of last resort; it evaporates the quarter one hyperscaler throttles capex, and with the same collateral potentially pledged twice, the true leverage stays unmappable until it crystallizes. Credit is the wire through which this reaches the real economy — the only precedents for synchronised equity-and-credit corrections are the GFC and March 2020. Equity is pricing the upside; credit is pricing the exit — and the credit line has not finished moving.