Applied Materials answered the week's central question Thursday night with the kind of report that closes debates. Record quarterly revenue, a Q4 guide of $10.25 billion that landed well above where the Street was parked, gross margins expanding for the 13th consecutive quarter, and CEO Gary Dickerson raising the calendar 2026 wafer fabrication equipment growth call above 30% while explicitly flagging "another strong growth year" for 2027. The AI buildout is not stalling in the equipment layer. Then Friday arrived, and AVGO dropped roughly 6% after Bank of America published a note flagging $370 billion in potential AI vehicle debt tied to Broadcom's custom-accelerator financing program. The stock had just proven, three months ago, that its AI chip revenue grew 143% year-over-year. The selloff was not a demand story. It was a balance-sheet anxiety attack. Those are different things, and the market spent Friday afternoon deciding whether it cared about the distinction.

The equipment layer just confirmed the AI buildout is real and accelerating. What's being argued now is who captures the next dollar of it, and on what terms.

AMAT Closes the Debate. Friday Opens a New One.

The fund context we wrote last Wednesday titled "CPI, Then AMAT Decides" was straightforward: the July inflation print mattered for the rate regime, and AMAT's report mattered for the AI supply chain thesis. Both delivered answers.

On CPI: Persistent above-target inflation, with core PCE running near 3.3% and headline readings elevated by energy prices and tariffs, continues to anchor the rate regime. The July number (Aug 12) came in at 0.1% month-over-month, 3.4% year-over-year headline, core at 2.5%, both annual rates ticked down, which is the right direction, but not the magnitude the market needs to price a cut. The Federal Reserve has held the federal funds rate steady at 3.50-3.75% for the fifth meeting in a row. Prediction markets now price roughly a 71% probability that the Fed maintains rates at the September meeting, with a 29% probability of a hike. That is the regime: not cuts-because-inflation-died, but hold-or-maybe-hike-because-it-hasn't. The rate read matters for **BAC** and for every multiple in the book. We are not in a rate-cut story. We are in a quality-earnings story, which is a different setup entirely.

On AMAT (Thursday, Aug 13): Applied Materials posted record Q3 results as AI-driven demand lifted revenues, with advanced packaging and memory growth. The company achieved its 13th consecutive quarter of year-over-year gross margin expansion. But the headline that matters most is the Q4 guide. Management guided for $10.25 billion in revenue, plus or minus $500 million, and $4.02 in non-GAAP EPS, plus or minus $0.20. At the midpoint, revenue would rise another 12% sequentially after Applied Materials just delivered the highest sequential revenue growth in its history. CEO Gary Dickerson also raised the company's calendar 2026 Semiconductor Systems outlook and said stronger customer visibility points to "another strong growth year" in 2027. That is not an equipment company hedging its forward visibility. That is a company with a full order book telling you exactly what it sees.

The non-consensus read: AMAT's guide is the cleanest leading indicator the AI supply chain has produced this cycle, more reliable as a demand signal than any hyperscaler capex announcement, because it reflects actual tool orders with committed delivery schedules, not CFO-approved spending guidance that can be revised in the next quarter. Applied Materials is the largest US maker of semiconductor equipment, and almost every major chipmaker, including the factories that produce Nvidia's AI processors and the memory chips alongside them, relies on Applied Materials' tools. When Dickerson raises the WFE growth call and adds explicit 2027 commentary, the order book behind that statement is already partially booked. This is not optimism. It is backlog math.

One asterisk the market is right to watch: free cash flow collapsed 80% to $210 million against record revenue. Applied Materials is investing heavily in capacity to support longer-term demand, including plans to double quarterly system manufacturing capacity by 2028. As of July 26, the company had $9.4 billion in cash and short-term investments. The cash is there. The FCF compression is a capital-investment story, not a deteriorating-business story. But it will remain a short's best argument until the capacity spend reverses into operating leverage, which is what 2027 is supposed to deliver.

AVGO Takes the Hit. AMD Gets the Upgrade. NVDA Waits Eleven Days.

Friday's most consequential price action came from the AVGO-AMD split, and the framing matters. AVGO dropped roughly 6% as Bank of America flagged $370 billion in potential AI vehicle debt, while AMD surged approximately 4% on Baird's Street-high $1,250 price target.

The AVGO story deserves precision. The $370 billion is not Broadcom's debt; the vehicle raises capital and leases custom AI accelerators to customers. Broadcom's fiscal Q2 2026 revenue was $22.19 billion, up 47.9% year over year, with AI chip revenue up 143%, and management has guided to $16 billion of AI semiconductor revenue for the current quarter. Polymarket traders gave Broadcom a 94% chance of topping $15 billion in AI revenue this quarter and a 78% chance of exceeding $16 billion, and Friday's selloff targets the financing structure rather than underlying demand. We own AVGO as part of the book's core AI-chip exposure. Friday's drop is a price disconnect, not a thesis break. The question the market is actually asking is whether the new era of AI infrastructure financing, SPVs, off-balance-sheet leasing structures, trillion-dollar private capital partnerships, creates systemic risk or simply represents the privatization of what would otherwise be government-scale buildout spending. That is a legitimate debate, but it is not a debate about whether Broadcom's custom silicon business is working. It clearly is.

Note also: Google has assembled a $200 billion financing program for Anthropic, with more than $150 billion tied to TPUs and involving Broadcom, Blackstone, Apollo, Morgan Stanley, and others. The NVDA $500 billion platform announced last week and the Google-Anthropic structure this week are the same animal: permanent institutional financing for AI infrastructure. The headline on AVGO was "debt risk." The actual read is that hyperscalers and their chip partners have figured out how to fund decade-scale buildouts without waiting for annual budget cycles. That is structurally positive for every company in our power and AI pillar.

On AMD: Baird analyst Tristan Gerra doubled his AMD price target to a Street-high $1,250 from $625 while maintaining Outperform, with his model implying AMD AI GPU platform revenue reaching $147 billion by 2030 on 15% share of the data center AI accelerator TAM. AMD had initially fallen after its August 4 Q2 report despite revenue of $11.54 billion, up 50.1% year over year, with Data Center revenue more than doubling to $6.72 billion; flat gross margin guidance of 56% for Q3 was the culprit, contributing to a post-earnings drop of roughly 9%. We own AMD as part of the AI chips pillar. The sell-off on the margin guide, the subsequent rally on Baird's target, this is the right kind of volatility around a name where the thesis is intact and the data center ramp is measurable and documented. The gross margin pressure is a product-mix and manufacturing-ramp story that resolves as MI300X and next-generation accelerator yields improve. Friday's move is the market recalibrating around that timeline.

And then there is NVDA. NVDA is set to report second-quarter earnings after the market closes on August 26, and Bank of America, maintaining its Buy rating and $350 price objective, expects the chipmaker to post revenue of $94 billion to $95 billion, a $3 billion to $4 billion beat above the company's own $91 billion guidance. Nvidia investors have not had the success they are accustomed to in 2026: the stock easily crushed the market in the first half of recent years, but has barely made investors money in 2026, rising about 2% for the year while the S&P 500 is up around 7%. We own NVDA as the anchor of the AI chips pillar and our conviction has not moved. The stock's YTD underperformance against its own earnings trajectory is the most important price disconnect in our book right now. When a company running 85-90% revenue growth trades at the same forward multiple as the S&P 500, either the multiple is going to expand or the growth is going to slow. AMAT's Q4 guide, Dickerson's 2027 commentary, BofA's revenue-beat setup, none of that points to slowing growth. August 26 is the event of the month, and then some.

The Rate Regime, BAC, and the Pillar That Keeps Working Quietly

The week's market internals told a familiar August story: consolidation after a strong recovery, not distribution. The Nasdaq 100 is on track for modest weekly gains in what is best characterized as healthy consolidation above the 50-day SMA following a strong bounce off the July 29 lows, while the S&P 500 and Russell 2000 notched fresh all-time highs. Stocks continued to melt up with market breadth expanding, with S&P breadth improving to above 75% of members trading above their 200-day moving average, above January highs. Breadth this broad, at all-time highs, is not a warning sign. It is the market telling you the earnings cycle is real and the participation is genuine.

The rate picture, with resurgent inflation tied to rising energy prices prompting some forecasters to expect higher rates before year-end and Fed Chair Kevin Warsh pledging a return to the 2% inflation target while offering few clues about his outlook, is the backdrop against which **BAC** operates. A bank running net interest income momentum in a hold-or-hike regime is a well-positioned bank. We hold BAC as the fund's deliberately uncorrelated large-cap financial. The thesis is simple and the rate regime is cooperating with it.

On MP Materials: the rare-earth and magnet reshoring story has been quiet newsflow-wise this week, but the structural thesis got stronger. Every AI financing structure announced this week, NVDA's $500 billion platform, Google's $200 billion Anthropic program, represents demand for data center hardware that ultimately requires permanent magnets for motors, cooling systems, and robotics. The DoD price floor and Apple offtake underpinning MP's domestic magnet supply chain are defense-industrial policy meeting a physical-infrastructure need that is, if anything, accelerating. MP is a position of ours. Quiet weeks are good weeks for a long-duration reshoring thesis.

On LEU: the enrichment reshoring story continues to digest. No material newsflow this week, and the stock is in repair mode. The DOE award underwrites the thesis. We treat this as a prove-it hold, not a conviction-size position. The catalyst is the next round of DOE contract disclosures and any update on HALEU program timelines.

On LLY: our GLP-1 position continues to compound as an off-AI diversifier. No earnings this week, but the metabolic-drug thesis is structural and the pipeline runway remains intact. Q3 earnings will be the next significant event for LLY holders.

What to Watch the Week of August 18

Two events stand above everything else on this calendar. PANW on August 21 will either confirm or complicate our AI software and security thesis. The setup coming in is extension-driven; the cohort got punished on beats this season when multiples ran ahead of earnings, and PANW is not cheap. What matters is whether the platformization narrative (customers consolidating onto PANW from point solutions) is accelerating or plateauing. CEO Nikesh Arora's commentary on AI security workloads will be the tell.

Jackson Hole (Aug 22): Warsh's posture heading into the September 15-16 FOMC will define the rate regime for the next quarter. At the Fed's June meeting, Warsh declined to submit individual economic projections, although nearly half of policymakers said they would support a rate hike later this year. If Jackson Hole brings a hawkish lean, the September hike probability (currently ~29%) climbs and BAC's NII thesis strengthens while growth multiples get another test. If Warsh signals patience, September stays a hold and the setup into NVDA's report becomes cleaner.

And then August 26. Nvidia faces questions about whether it can break its four-quarter pattern of post-earnings underperformance relative to expectations. BofA's $94-95 billion revenue estimate implies a beat of $3-4 billion over company guidance. AMAT's Q4 guide and Dickerson's 2027 commentary are the most recent supply-chain confirmation that Blackwell demand is running ahead of prior models. The stock's YTD underperformance sets up the asymmetry: if NVDA beats and guides up, the gap between business performance and stock price closes. If it meets and guides in line, some of the current multiple case gets tested. We are not managing around that event. We are positioned through it.

Forward Thesis by Pillar

AI Chips / Hardware (NVDA, AMD, AVGO, MU): Conviction intact, elevated into August 26. AMAT's Q4 guide of $10.25 billion and the 2027 growth commentary from Dickerson represent the most concrete supply-chain validation of the cycle. NVDA at 2% YTD against that business backdrop is the fund's highest-conviction price disconnect. AVGO's Friday drop is noise against 143% AI revenue growth; the Q4 guide and September 1 earnings call are what changes the narrative, if anything does. AMD's margin pressure is a known, dated, and resolving issue. What would change conviction: NVDA guiding Q3 in-line or below on August 26, or a credible signal that hyperscaler capex is being pulled forward rather than sustained.

Power & Grid (GEV, VST, CEG): The "firm power and the wires, not the panels" thesis continues to be the quiet winner of the summer. Every AI financing structure announced this week represents demand for 24/7 power and the grid infrastructure to deliver it. GEV sits near highs on backlog momentum; CEG has repaired strongly. No single catalyst this week, but the structural bid for firm power names is getting stronger as the financing architecture for the buildout becomes more permanent and visible. Quanta (PWR) remains the grid-layer name we are watching most closely as the get-paid-regardless expression of the wires thesis.

AI Software / Security (PANW, NET, NOW): PANW earnings August 21 is the near-term test. NET's edge-inference and AI-agent traffic story keeps compounding quietly; this is a name we hold with genuine long-term conviction. The risk in the software cohort is multiple compression on beats rather than earnings misses, which is a posture risk, not a thesis risk. We treat any post-PANW-earnings weakness as a thesis check rather than a trigger.

Critical Minerals / Reshoring (MP): Long-term conviction, patient timeline. The DoD floor and Apple offtake create an asymmetric risk profile for MP's domestic magnet supply chain. The AI infrastructure financing boom announced this week is the demand signal; the supply chain for the physical hardware that runs AI is exactly what MP services. A position of ours. No urgency to change the view.

Fuel Cycle / Energy Security (LEU): Prove-it hold. The DOE award is the underwrite; the catalyst is the next round of contract disclosures. Quiet week was consistent with the thesis; no thesis change.

GLP-1 / Healthcare (LLY): Off-AI diversifier, structural position. The metabolic drug thesis is intact; Q3 earnings will be the next meaningful check on pipeline execution. We hold LLY with conviction as a deliberate diversifier from the concentration in AI and power.

Financials (BAC): Resurgent inflation tied to rising energy prices has prompted some forecasters to expect higher rates before year-end. That is the tailwind for BAC's net interest income trajectory. A hold-or-hike regime is a good regime for a well-capitalized large-cap bank. Warsh's Jackson Hole tone next week is the next real catalyst for the rate pillar.

Nova One Capital and the Nova X Convergence Fund hold positions in securities discussed in this letter, including NVDA, AMD, AVGO, MU, GEV, VST, CEG, PANW, NET, NOW, MP, LEU, LLY, and BAC. This disclosure is provided as standing context so readers can follow the fund's thinking. Nothing in this letter constitutes a current trade signal, allocation disclosure, or investment advice.

This document is for informational purposes only and does not constitute investment advice, a solicitation, or an offer to buy or sell any security. The views expressed are those of Nova One Capital and the Nova X Convergence Fund as of August 15, 2026, 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. Nova One Capital and the Nova X Convergence Fund hold positions in securities discussed herein. Readers should conduct their own due diligence and consult a qualified financial advisor before making any investment decisions.