AI Chip Cycle Confirms Durable Demand

July Opens with AI Confirmed, Energy Deflating

The first session of the second half of 2026 delivered a clear verdict: technology sector leadership is not unwinding — it is consolidating on the back of proven earnings power. The Technology Select Sector SPDR (XLK) gained +2.76%, QQQ added +1.70%, and the S&P 500 rose +0.78% in a tape where the dominant driver was backward-looking confirmation rather than forward speculation. The headline data point defining today's session is Q2's record chip rally: Micron, Intel, and AMD collectively added $2 trillion in combined market capitalization during the quarter. That is not narrative. That is the AI infrastructure demand cycle producing one of the most significant semiconductor value-creation episodes on record, and the market is now pricing it as a durable earnings cycle rather than a positioning event.

The other signal came from oil. Iran confirmed it is exporting crude at a 20% premium as the US-Iran MOU lifts the blockade, and June closes as oil's biggest monthly decline. The Hormuz tail risk that entered portfolio analysis in late May has been substantially priced out. What remained ambiguous going into today — whether Iran would sustain exports under economic pressure — was answered with a pragmatic data point: Tehran is selling, and selling at a premium to spot. The geopolitical risk premium that inflated energy-linked inflation expectations has compressed materially.

The one complication is LNG. QatarEnergy extended its force majeure declaration through September for Italian shipments, a reminder that energy supply normalization is multi-layered. This is a Europe-specific issue and does not affect the domestic US thesis, but it introduces residual volatility in European natural gas markets that could feed back into global sentiment if conditions deteriorate further.

Portfolio Changes: AMZN Trimmed, NVDA Raised

Two trades execute today, both driven by price action evidence rather than thesis revision.

AMZN trimmed from 20% to 17%. A -0.75% session in a tape where XLK gained +2.76% and QQQ added +1.70% is a meaningful underperformance signal. The position was raised to 20% on June 30 on the argument that recovery was in progress and 20-day momentum was approaching the flat line. Today, the 20-day slipped back to -1.25%. When a position described as recovery-in-progress declines on the strongest tech day in weeks, the recovery is not confirmed. Pershing Square's 17.4% and Bridgewater's 4.1% institutional weights anchor the thesis, and AWS geopolitical tailwinds are real — but position sizing follows what the stock is doing, not what the thesis projects.

NVDA raised from 10% to 13%. The +2.63% session outperformed QQQ by nearly a full percentage point on the strongest tech tape in recent memory. The 20-day momentum has improved from -5.79% to -2.75% and the 60-day from -5.44% to -3.14% — compressing across both timeframes. The Q2 semiconductor cycle confirmation makes the momentum improvement signal meaningful rather than random: Blackwell GPU demand is being validated not by management commentary but by competitor order books and downstream chipmaker revenue. Bridgewater holds NVDA at 3.7%. The raise to 13% reflects conviction in a directional inflection, not a full reversal call on a name where momentum remains technically negative.

Institutional Signals

Bridgewater's book (SPY 12.7%, IVV 10.7%, AMZN 4.1%, NVDA 3.7%) continues to read as a vote for broad market participation anchored in AI-exposed equities — consistent with the QQQ-led tape environment of the past several weeks. Pershing Square's concentration in AMZN (17.4%) and MSFT (15.3%) provides the conviction floor for both positions despite their continued negative intermediate-term momentum; Ackman's time horizon extends beyond current price action. Berkshire's absence from the technology-AI complex is a known stylistic constraint; BRK-B gained +0.89% today and holds consistently positive momentum among the non-LLY, non-IWM candidates, a sign that Buffett's conglomerate is compounding quietly in the background without requiring exposure to the AI narrative.

Michael Burry's disclosed short on Caterpillar — after it nearly doubled in the AI-driven rally of 2026 — is the session's most instructive institutional signal. Burry is targeting infrastructure derivative beneficiaries of the AI trade rather than the AI names themselves, suggesting his concern is not about AI demand but about stretched valuations in adjacent capital goods equities. That framing is consistent with holding NVDA as a direct infrastructure beneficiary while declining to add XLI as an infrastructure derivative.

What Could Break the Thesis

Iran deal deterioration. The MOU is producing oil exports but has not been formally codified. QatarEnergy's force majeure extension shows that even within a broadly stabilizing environment, individual supply participants can still disrupt flows. A collapse in the US-Iran framework would rapidly reverse oil's monthly decline and re-inject inflation into rate expectations, pressuring the growth equity multiples that underpin this book.

AI ROI failure at the enterprise layer. CNBC reported that employers who laid off workers citing AI are already reversing those decisions — a nuanced signal. It could indicate that AI productivity is falling short of initial promises, which would eventually pressure hyperscaler capex commitments. AWS, Azure, and Google Cloud are building infrastructure on the assumption that enterprise AI deployment will absorb capacity at scale. If enterprise adoption stalls or proves economically insufficient, the investment cycle could reverse before supply does.

Fiscal-driven rate pressure. TLT declined -1.18% today despite oil's monthly decline removing a key inflation input — an unusual divergence. Long yields rising on fiscal rather than inflation grounds would compress growth equity multiples across the entire book independently of earnings quality. The portfolio's growth tilt makes it structurally sensitive to this scenario, and the mechanism would not be solved by rotating into defensives without fundamentally changing the character of the book.