The market handed us a five-act drama this week, and by Friday's close, the AI infrastructure thesis had won every scene that mattered. Nasdaq +1.74% for the week, S&P 500 +1.23%, with the Dow actually down 0.50% on Hormuz-driven energy panic. Growth beat value by the widest margin in weeks. Information technology led all eleven S&P sectors. A $30 billion Apple–AVGO chip deal landed Wednesday. META ripped more than 6% Friday on reports it is entering cloud computing using surplus AI capacity and building a new custom AI chip alongside AVGO. And SK Hynix's $26.5 billion U.S. IPO, the largest in recent memory for the sector, debuted more than 17% above its offer price. That is not a market losing faith in the AI supercycle. That is a market repricing it higher through the noise.

"The Hormuz spike, the Fed minutes, and the Samsung miss were all sentiment events. The Apple–AVGO deal and META's cloud pivot were structural ones. This week's tape separated signal from noise more cleanly than any week in Q3 so far."

The Week That Clarified the Thesis

Here is the non-consensus read: this week's volatility, in hindsight, was a compression spring. The early-week selloff, Samsung disappointing against impossibly high expectations, Hormuz re-escalating, the Fed minutes flagging a "hawkish bias", shook out the momentum tourists. What remained by Friday was structural confirmation after structural confirmation of exactly what Nova X was built to own.

Start with the Samsung noise. Samsung reported a 19-fold increase in profit and still got punished. The best-performing corner of the stock market was pummeled, with the semiconductor gauge sinking over 4.5%, not even a record profit was enough to entice investors. Our read then, and now: this is the tax on being the consensus trade. The bar for semis is priced at extraordinary; ordinary extraordinary no longer clears it. But the demand destroying Samsung's stock narrative is not in question, it drove the rebound two sessions later. AVGO advanced 3.20% on Thursday after confirming the renewed chip supply agreement with Apple through 2031, in a deal exceeding $30 billion covering wireless filters and custom AI silicon. The Samsung event was a bar problem. The Apple deal is a contract, multi-year, domestic, signed.

Then the geopolitical shock. The Dow fell sharply Wednesday after President Trump told the NATO summit in Turkey that the ceasefire with Iran is "over" amid renewed hostilities that sent oil prices surging. WTI futures rose 4.4% to $73.52 per barrel; Brent jumped 5.2% to $78.02. But here is the tell: while the blue-chip Dow lost 576 points on Wednesday, the Nasdaq actually rose 0.2%. The market knows the difference between cyclical oil-shock pain and AI infrastructure compounding. Energy pain is a cyclicals story. The AI capex supercycle is a decade-long structural story. Hormuz does not cancel NVDA's H100 backlog.

By Friday, the synthesis was visible. META led S&P 500 stocks, surging more than 6% on reports that it plans to enter the cloud computing market using its excess AI capacity and build a new AI chip in partnership with AVGO. NVDA gained 4.0% to rank among the top Dow performers. AVGO was up 11% for the week, helped by the Apple confirmation. And the Nasdaq finished with a 1.74% weekly gain to 26,281; the S&P 500 rose 1.23% to 7,575. The thesis did not just survive the week. It accelerated.

Pillar 1 · AI Chips / Hardware: AVGO's Apple Deal Is a Decade-Long Contract, Not a Trade

The week's single most consequential catalyst for the chip pillar was not a quarterly print or an analyst upgrade, it was a procurement contract. Apple's $30-billion-plus commitment to AVGO through 2031 covers wireless filters and custom AI silicon. Broadcom followed the announcement by confirming plans to expand its U.S. facility. Evercore ISI immediately called it a strategic positive. One analyst put a $600 price target on AVGO, and given that AVGO's market value sits within about 5% of the $2 trillion club, with AI semiconductor revenue up 143% year over year last quarter, that call does not seem heroic.

The AVGO thesis implication: the Apple deal is exactly the kind of locked-in, multi-year, domestic-supply-chain-anchored revenue stream that makes AVGO categorically different from commodity chip exposure. This is not cyclical DRAM. This is contracted custom silicon with the most cash-generative consumer technology company on the planet as the counterparty. The valuation question becomes less about near-term multiple compression and more about how many additional hyperscaler or hyperscale-adjacent customers follow Apple's playbook. OpenAI already has. Meta appears to be next.

On the rest of the pillar: ARM Holdings closed Thursday at $327.87, up $27.75, as UBS raised its price target to $470 and TD Cowen lifted its target to $475, both maintaining Buy ratings and citing strong demand in AI and cloud computing ahead of ARM's fiscal Q1 earnings due July 29. DELL added 4.22% Thursday after Evercore ISI raised its price target to $500; Dell is sitting on a $51.3 billion AI server backlog with approximately $60 billion in AI server sales guided for the full year. The infrastructure layer is not slowing. The backlogs are growing.

Pillar 2 · Nuclear / Power Infrastructure: Hormuz Spike Proves the Baseload Premium

The Iran escalation this week was, counterintuitively, a constructive data point for the nuclear and power pillar. Here is the logic: the end of the U.S.-Iran ceasefire is driving oil prices higher; on Tuesday and again Wednesday, the U.S. said it was striking Iran in retaliation for attacks on commercial ships in the Strait of Hormuz. WTI was trading near $75 per barrel by Wednesday evening, up from roughly $69 at the start of the week. While elevated, prices remain well below peaks seen at the height of the war, when they neared $120 per barrel.

The non-consensus read: every Hormuz spike is a reminder that fossil fuel energy has a geopolitical tail risk that baseload nuclear simply does not carry. VST and CEG do not import oil through a contested strait. They generate power from plants on U.S. soil, under long-term PPAs, increasingly backstopped by AI data center demand that itself is indifferent to Middle East maritime traffic. The energy risk premium that Hormuz reintroduces into the market accrues to reliable baseload generators, not to refiners, shippers, or spot-price exposed utilities.

The Fed minutes add a second layer. The June FOMC minutes stressed the risk of higher inflation; the minutes demonstrated the Fed's "hawkish bias," highlighting that upside inflation risks remain while concerns around the labor market have eased. Phillips noted that inflationary pressures are not necessarily limited to energy, as cost increases are starting to show up in other areas including electronics. A higher-for-longer rate environment with energy risk embedded in oil markets is precisely the backdrop in which contracted baseload power with inflation-linked pricing clauses becomes more valuable, not less.

Pillar 4 · AI Software / Platforms: META Is No Longer Just an Ad Platform

Friday's META move deserves its own section because the market may be under-pricing what the cloud computing report actually signals. META surged more than 6% on reports that it plans to enter the cloud computing market using its excess AI capacity and that it wants to build a new AI chip in partnership with AVGO. DA Davidson analyst Gil Luria framed the opportunity cleanly: "Meta has been a significant beneficiary from the advances in AI by selling more ads at higher prices, which has driven significant revenue acceleration. It has not gotten credit because it has increased capex even more. If Meta slows down capex and starts monetizing it, we see significant upside to revenue and cash flow."

This is the thesis inflection that changes META's valuation conversation entirely. The market has been discounting META as a high-capex, single-revenue-stream business with an AI side project. If cloud compute monetization becomes real, selling excess GPU capacity to enterprises the way AWS monetized Amazon's internal infrastructure: META becomes a two-engine revenue machine. The AI ad monetization thesis is already intact and compounding. A cloud revenue layer on top of that would be a re-rating catalyst, not an incremental upgrade.

PLTR is quiet this week but holds its position as the enterprise AI OS thesis anchor ahead of the broader Q2 earnings season. The PLTR print is not this week, but the setup, government AI contracts accelerating, commercial AIP adoption broadening, has not changed. We are watching for any incremental contract announcements through July.

Pillar 3 · Digital Finance Rails: Still Waiting for the Catalyst

No material stablecoin or institutional crypto legislative catalyst crossed the tape this week. Bitcoin hovered near $63,000-$64,000 per the Yahoo Finance data. The GENIUS Act / CLARITY Act legislative calendar remains the master unlock for this pillar, and it was not this week's story. We are watching, not chasing. The Digital Rails pillar remains flat until a clean 30%+ entry presents itself against a confirmed legislative catalyst. Discipline here is the thesis, not passivity.

The Calendar That Runs the Show

Next week is the most consequential macro and earnings calendar of Q3 so far. Three categories matter to the fund's pillars.

Macro: June CPI and PPI. The June CPI release is scheduled for Tuesday, July 14, at 8:30 a.m. ET. This is the primary inflation input for the July 28-29 FOMC meeting. Context: the May CPI showed headline inflation rising 0.5% on a seasonally adjusted basis, up 4.2% over the last 12 months. Core CPI showed recent moderation, with the May reading at 0.2% month-over-month and 2.9% year-over-year. The Hormuz oil spike this week introduces upside risk to the June headline number. If energy prices bled into the June print, the hawkish Fed narrative from the minutes gets reinforced and the July FOMC hike probability, currently priced at approximately 25% by CME FedWatch, moves higher. That is the primary macro risk to tech multiples next week.

Earnings: Big banks + semiconductor OEMs. JPMorgan Chase, Bank of America, Citigroup, Goldman Sachs, and Wells Fargo are scheduled to report Q2 earnings before the bell on Tuesday. Bank earnings set the macro tone for the rest of the week. Strong results with net interest income holding will signal that the rate environment is manageable; guidance deterioration on credit quality would reignite recession concerns. ASML and Taiwan Semiconductor Manufacturing (TSM) are highlights for the semiconductor sector next week. TSM's report is the most direct read on AI chip demand that exists in public markets: NVDA, AMD, and AVGO all run through TSMC's fabs. A strong TSM print with AI revenue commentary would be the most potent confirmation catalyst for the chip pillar heading into the late-July hyperscaler capex season.

The levels that matter: the Nasdaq closed Friday at 26,281. The S&P at 7,575. A hot CPI print Tuesday that pushes the July hike probability above 40% is the primary catalyst for a tech multiple compression event. A TSM print with strong AI revenue guidance would be the countervailing force, and based on every demand signal this week (Apple's $30B AVGO commitment, META's cloud pivot, SK Hynix's blockbuster IPO), we expect TSM to deliver. The setup going into next week is fundamentally constructive; the tail risk is a CPI that hands Warsh's hawkish wing exactly the data it is looking for.

Conviction and What Would Change It

Pillar 1 · AI Chips / Hardware (NVDA, AMD, AVGO): Conviction HIGH and rising. The Apple–AVGO deal is the week's structural signal. A multi-year, $30B+, U.S.-domestic manufacturing commitment from the most cash-generative consumer tech company in the world is not a trading position, it is a supply chain decision with a ten-year horizon. The META custom chip report Friday adds a second hyperscale-adjacent customer to the custom silicon thesis. NVDA remains the infrastructure backbone: every AI system being built, including META's cloud compute offering and AVGO's custom AI chips, requires NVDA's CUDA ecosystem either directly or as a benchmark. What would change conviction: a TSM earnings miss with explicit AI demand deceleration commentary, or NVDA guidance cut at the late-July capex checkpoint. Neither is our base case.

Pillar 2 · Nuclear / Power Infrastructure (VST, CEG, LEU): Conviction STEADY, with the Hormuz spike as an additive data point, not a threat. The thesis has never required oil to stay cheap, it requires AI data center power demand to keep growing, and that demand is not correlated to Strait of Hormuz traffic. The LEU DOE 2-year option recheck around mid-July is the next binary for this pillar. A confirmed DOE exercise is a clean hold signal; an explicit non-renewal would trigger a thesis-based reassessment, not an automatic exit. The commercial HALEU path and monopoly moat survive the DOE contract regardless, the sizing is already calibrated to that optionality structure.

Pillar 3 · Digital Finance Rails: Conviction MUTED, pending legislative catalyst. No GENIUS Act / CLARITY Act movement this week. Bitcoin near $63,000-$64,000 is range-bound. We are not chasing this pillar without a clean entry against a confirmed stablecoin legislative unlock. The discipline of waiting is the thesis, not the absence of one.

Pillar 4 · AI Software / Platforms (META, PLTR): Conviction HIGH on META, RISING. The cloud compute monetization report changes the investment case structurally. DA Davidson frames it precisely: META has driven significant revenue acceleration from AI-powered ad pricing, but "has not gotten credit because it has increased capex even more." If cloud compute becomes a second revenue engine, the capex narrative inverts from a headwind to a moat. Q2 earnings (late July) is the near-term confirmation event. PLTR conviction remains intact ahead of its own earnings, the government AI contract acceleration has not reversed, and the commercial AIP pipeline broadening is the Q2 story to watch.

Pillar 5 · Specialized GPU Cloud (CRWV): Conviction FRAGILE, stop-watched. CRWV remains the most execution-dependent name in the book. No material news this week. The stop is the line in the sand. The AI-infra demand thesis that backstops CRWV, contracted GPU cloud, NVIDIA-aligned, is intact at the macro level; the single-name risk is operational, not structural. We need a clean week with no AI-infra demand signal reversal to hold conviction here.

The master checkpoint that governs everything: traders are pricing in a Fed hike as early as September, then see policymakers staying on hold for at least the next year. Bank of America has raised its interest rate forecast, now seeing three quarter-percentage-point hikes before year-end. That is the tail that the CPI print on Tuesday is either going to validate or defuse. The July 14 CPI release carries significant weight in the context of the FOMC's July 29 rate decision. If headline CPI re-accelerates on energy passthrough from the Hormuz spike, the fund's late-July hyperscaler capex checkpoint becomes even more critical as the fundamental offset to rate-hike multiple compression. The thesis holds in that scenario: AI capex is not rate-sensitive at the hyperscaler scale, but the price action would be volatile, and that volatility is where discipline earns its keep.

This document is for informational purposes only and does not constitute investment advice or a solicitation to buy or sell any security. The Nova X Convergence Fund letters reflect the views and analysis of the fund's research team as of the date of publication and are subject to change without notice. Past performance is not indicative of future results. All investments involve risk, including the possible loss of principal. Nothing herein should be construed as a recommendation to take any specific investment action. Published July 11, 2026.