Silicon’s Bull, Five Layers Deep
The semiconductor leadership extends beyond the megacaps — interconnects, memory, power, packaging, materials all leading.
Edition · April 26, 2026 · KW 17
Thank you for reading this week’s edition of Closelook@US Stock Markets, dated April 26, 2026 👋. The next edition will be published on May 3, 2026.
For the schedule of our other publications, please check the publishing schedule at the end.
Table of Contents
1. This Week’s Action
2. Overview: The US Market Map
3. The State
4. The Outlook
5. What May Lie Ahead: Macro Setup
6. What May Lie Ahead: Technical Setup
7. The AI Build-Out Portfolio: What We Did
8. The AI Build-Out Portfolio: What We Plan To Do
9. Knowledge Corner
10. Upcoming Transactions: Be Informed
11. What May Go Wrong: Risk & Change Triggers
12. Final Words
(1) This Week’s Action
New all-time highs on S&P, Nasdaq-100, and Nasdaq Composite. SOXX printed an 18-session winning streak — a record since the ETF’s 2001 inception. ServiceNow lost 17.75% in one day on a headline beat — but IGV at the index level held flat on the week.
The damage was idiosyncratic. The earnings cluster of April 27 to May 1 delivers the structural verdict.
The numbers above set the structural read at a glance. The headline: semis vertical (SOXX +11.04%, SMH +9.11%), Rubin Build-Out 100 extending its YTD lead at +65% (EW and MW), software at the index level held flat (IGV +0.14%) despite NOW and IBM cracking individually. The chart below shows where the Nasdaq-100 sits on the long-term log channel.
The four timeframes capture the same trend at four resolutions. The next table summarises the indicator readings on the weekly view.
The macro indicator picture is constructive. The diagnostic read shifts when we cut the index by composition rather than by time — the chart below decomposes QQQ into Tech vs. Ex-Tech and Top-20 US stocks (not confined to the Nasdaq-100) vs. Tail-70 to show exactly what is driving the index this week.
What Happened
The major US index proxies closed the week at fresh territory. S&P 500 at 7,165.08 Friday, up +0.55% on the week to a fresh all-time high. Nasdaq-100 (NDX) at 27,303.67, up +2.37% on the week — QQQ tracked it at 663.88, +2.32% on the week from the prior Friday’s 648.85.
The broader Nasdaq Composite at 24,836.60, +1.50% on the week. Dow Jones at 49,230.71 was the lone diverger, finishing −0.41% on the week. IWM (Russell 2000 ETF) at 276.65 added +0.32% on the week — the broader Russell 2000 index closed at 2,787.00. VIX rose 7.04% on the week to 18.71 — volatility creeping back from the Hormuz-driven low but still historically benign.
The headline within the headline was the semiconductor extension. SMH closed at 506.44, +9.11% on the week. SOXX closed at 461.60, +11.04% on the week — SOXX printing its 18th consecutive up day on Friday, the longest streak in the ETF’s history since its 2001 inception. Per a NYSE research note circulated mid-week, semiconductors now account for ~15.5% of the S&P 500’s market-cap weight and have contributed roughly 4.9 percentage points of the index’s 12.8% rally since March 30 — about 40% of the entire move.
Two single-name prints framed the week. Intel surged +20.50% on a beat-and-raise, finally closing above its dot-com 2000 high — a 26-year price ceiling, finally cleared. Nvidia +3.27% on the week to retake the $5 trillion market cap mark on Friday’s session.
And then there was the other side of the tape. ServiceNow lost 17.75% on Wednesday, on a headline beat. Revenue beat consensus, cRPO up 25% YoY, the Now Assist ACV target raised from $1.0B to $1.5B. The market cap fell from $107B to $87.8B in a single session—roughly $20 billion evaporated.
IBM lost 6% the same day, also on a revenue and EPS beat with raised software guidance. Two beat-and-lose prints on the same day are not a coincidence. They are a market that is repricing a structural equation in real time.
We covered the bridge math in Friday’s Daily Pulse — legacy decay + insufficient AI uplift + hybrid-instead-of-outcome pricing = premium multiple no longer defensible.
The diagnostic feature: IGV closed Friday at $85.20, essentially flat on the week (+0.14%). The two single-name beats-and-losses cracked NOW (W/W −6.71% to $90.17) and IBM (W/W −8.48% to $231.98) — but the broader software complex held. The Friday +5% bounce on NOW alone reversed most of the week’s mid-week damage at the index level.
The IGV/QQQ ratio tested the 0.1255 trigger zone mid-week and recovered to close the week at 0.1284, above the line. The damage was idiosyncratic, not sector-wide. The structural hold level at $74.62 sits more than $10 below the current IGV close. Software did not crack — two specific names did, and the rest of the cohort absorbed it.
Two Friday signals tilt encouraging. NOW closed Friday at +5% from the Wednesday capitulation low — early evidence that the bad news may already be priced in. SAP, Europe’s enterprise software bellwether, printed positive earnings on Friday and held its bid. Neither resolves the verdict. Both are the first data points in the direction of a stabilization rather than a continuation lower.
The earnings cluster of April 27 through May 1 will deliver the resolving signal.
That contrast is not a contradiction. It is the rotation framework working — and the question of whether the software bottom is in remains open through Friday.
Six of the top ten weekly winners across the entire S&P 500 are Rubin Build-Out 100 constituents. AMD, TXN, INTC, ON, MCHP, NXPI — semiconductors dominating the broad-index single-name leaderboard. The within-tech rotation thesis is visible at the SP500 level, not only inside the functional index.
Week Losers
The defense sector cracked. NOC and LMT both down >13% on the week. Coverage centered on election-cycle de-rating risk is being priced into the sector. Notably absent from the SP500 winners list are defense names that dominated in 2025 — a regime change worth watching.
Note: NOW (−6.71% W/W to $90.17) and IBM (−8.48% W/W to $231.98) sit just outside the top-10 losers despite their Wednesday capitulation moves. The Friday +5% recovery on NOW pulled it back from the deepest weekly drawdown.
(2) Overview: The US Market Map
Live constituents and full sector dispersion.
(3) The State
Silicon leadership defines this bull market — and the leadership runs five layers deep. SMH +9.1% on the week, SOXX printing an 18-session winning streak (a record since 2001), Intel +23% through its 26-year price ceiling, Nvidia retaking $5 trillion. Those are the visible names.
Underneath them, the AI hardware supply chain extends across the full build-out — The bull is wider than NVDA and INTC.
The Thesis — Silicon Leadership, Five Layers Deep
The semiconductor leadership that has carried this bull market since 2023 is broader than the headline names. The Rubin Build-Out 100 is up roughly 65% YTD on equal weight, ~70% on momentum weight — against Nasdaq ~6%. A 10× spread. That is not a quirk of construction. It is the framework that is working: leadership is focused on the AI hardware build-out, but distributed across the entire supply chain.
Inside the index this year and week:
Power Semiconductors +68% ytd on GB200 and Rubin rack architecture power-delivery — ON, MTD, STM, MPWR - all up +30 % this week
High-Speed Interconnects are up massively with Applied Optoelectronics, Lumentum, Marvell, and Coherent,t leading
Memory at +115% YTD — SK hynix, SanDi, SK, and Micron leading.
The Software Question — Verdict Pending
The April 19 Weekly Signal Special set up a litmus test for the most-attacked names in software. The setup ran exactly as forecast: IGV bounced double-digit off the March low — the expected revert-to-mean from oversold extremes. The bounce arrived on schedule, played out on schedule, and reached the levels where the second test would begin.
Then news arrived. Anthropic’s Claude Design launch as the proximate stress test for creative tools SaaS—ServiceNow’s earnings on Wednesday. The verdict on the bounce came mid-week: ServiceNow lost $20 billion of market cap in a single session — on a headline beat. IBM lost 6% on the same day, also on a beat.
Both prints reflected a market that has reached a working hypothesis on enterprise SaaS economics in the AI era: legacy decay + insufficient AI uplift + hybrid-instead-of-outcome monetization = premium multiple no longer defensible. The bridge math is in Friday’s Daily Pulse.
The earnings cluster of April 27 through May 1 carries the resolving signal—Microsoft on Tuesday with the Azure AI growth print. Alphabet, Meta, and Amazon are following. ServiceNow Analyst Day on May 4. Each one is a vote on whether the bridge math applies broadly across enterprise SaaS or stays idiosyncratic to NOW. Monday’s analyst reactions to the NOW/IBM prints will set the early tone — institutional desks return from the weekend with their re-rating, and the first hour of Monday’s tape is the leading indicator for the week.
The Frame — Within-Tech, Not Out of Tech
Two divergent tapes — silicon expanding across layers, software facing a verdict — are not a contradiction. They are the same thing observed from two angles. Both are the ninth within-tech rotation since the October 2022 low.
The pattern is documented and repeated: 2023 SVB, 2023 rate scare, 2024 sticky CPI, 2024 yen carry unwind, 2024 AI capex bubble wave one, January 2025 DeepSeek, March 2025 Liberation Day, summer 2025 AI capex bubble wave two, Q1 2026 equal-weight rotation. Every time, the tech selling concentrated in the previous cycle’s leaders while capital migrated to the next. Every time, the call for rotation OUT of tech was actually rotation WITHIN tech. This week is data point number nine. Software-as-a-service is the previous cycle’s leadership; AI hardware build-out is the next.
The Tech Thematic split is the diagnostic feature of the week — but not in the direction the headline narrative suggests. SOXX +11.04%, SMH +9.11%, IGV +0.14%. Software at the index level held flat; ARKK (- 3.45%) and CLOU (+0.10%) bracket the broader risk-software complex. CIBR +1.14% (Cybersecurity) and AIQ +1.92% (Broad AI) — both held positive.
The narrative “software cracked” is technically wrong at the index level. What cracked were two specific names: NOW −6.71% W/W and IBM −8.48% W/W. The AI trade extended into infrastructure (semis vertical) and held its software base at the index level.
Seasonality Setup
The seasonal map matches every other framework on the table. April 2026 is on track to print one of the strongest Aprils in the QTEC monthly history — currently around +18.54% month-to-date through April 23, the best April on record.
May seasonality thins (44% hit rate, average return barely positive). June is a coin flip. July is the destination — 4.1% average, 94% hit rate, the strongest seasonal window QQQ sees all year.
The regime reads Risk-On, with leadership concentrated in semiconductors and distributed across five build-out layers. Headline indices held or extended; thematics narrowed back into AI infrastructure but expanded within it. Breadth at the index level participated but did not lead; breadth inside the AI hardware supply chain was the diagnostic feature of the week.
Directional flow within the Rubin universe shows a continued handoff: Power Semis confirming the Mid-Ramp layer joining; Memory and Foundry stabilizing after weeks of distribution; Design heating up, signaling the next chip cycle is loading. This is the textbook signature of a within-tech rotation in progress, not a regime change at the index level.
Buildout Timeline — which stage leads the cycle?
The handoff phase continues. Design (Stage 1) heating signals the next chip cycle is loading — ARM, EDA names, and the architects of post-Rubin generations. Fabless / Foundry / Memory cooling Δ21 confirms that the current cycle is shipping and that those positions are rotating out — which aligns with the Hopper-era Sunset reading on the Generation Rotation Framework. Interconnect heating at the tail confirms back-end names are seeing the output flow.
The read pattern is consistent: the cycle leads the price. WDesignesign — if ARM and EDA names keep accelerating over the next 2-3 weeks while Memory and Foundry stabilize, the next leg is loaded 6-9 months ahead of the visible price move. That is the Δ21 signal in full operation.
(4) The Outlook
The Outlook rests on a Mag-7 earnings cluster across three days, the FOMC, and the ServiceNow Analyst Day. GOOGL Tuesday AMC. MSFT, META, AMZN, Wednesday AMC. AAPL Thursday AMC. FOMC Wednesday with SEP and Powell presser. ServiceNow Analyst Da,y May 4. Each one resolves a specific question.
Each one can extend, confirm, or invalidate the within-tech rotation read, and each one feeds into the verdict on the software bottom.
Forward Reads — Indices, Sectors, Thematics
QQQ leads SPY into next week on the Microsoft print. The 1W spread — XLK +3.80% vs the broad index — confirms tech is absorbing rotation dollars at the megacap level. IWM lags. Small-caps have no sector tailwind, no thematic catalyst, and the relative-strength story has not turned in their favor since the Q1 head-fake.
The index to watch is QQQ, which pulls away from SPY by 200+ basis points on any continuation day, driven by Mag-7 earnings.
Sector rotation tilts further into growth and away from defensives. XLK and XLY remain the twin engines into earnings. XLE caught the oil-rebound bid this week — sustainable only if the Hormuz situation re-escalates; otherwise, it means reversion to the mean. XLF holds its bid but trails meaningfully. Watch for XLC to accelerate on the Alphabet and Meta prints — Communication Services has been a base-building theme for three weeks, and platform-scale AI monetization needs a fundamental confirmation moment.
Thematic leadership has narrowed. SMH and SOXX have decoupled from IGV. Last week, these traded as one cohort; this week, they diverge by 10+ percentage points. The structural read is unchanged — AI infrastructure leads, software is on test. AIQ and CIBR are the names with 1M backing in the AI complex; their hold or break this week is the primary signal for the orchestration layer.
The Closelook indices to track this week. AW25 Agentic Winners is the index whose composition gets directly tested by next week’s Mag-7 earnings cluster. A held NOW $81.24 / IGV $74.62 floor — discussed below — keeps the AW25 setup alive for selective adds. A break of either kills it. The probabilistic case tilts toward the hold given Friday’s NOW +5% recovery and the SAP positive print, plus the multiple compression already 85-90% complete and NOW now trading below the S&P P/E for the first time in its post-IPO history. But the probabilistic case is a tilt, not a verdict. Monday’s tape and Tuesday’s Microsoft print are the first tests.
IGV $85.20 / QQQ $663.88 = ratio 0.1284. The ratio tested the 0.1255 trigger zone intraday on Wednesday and recovered to close above it. Bottom case for software needs continued holds above 0.1255 plus a reclaim of the prior-week 0.1311. Watch the next 5-10 trading days — that window contains more information than a normal quarter.
Watchlist — What Would Flip the Read
The next ten trading days (April 27 → May 8) deliver six resolving signals:
Monday April 27: analyst opinions on NOW/IBM. The first hour of US trading is the leading indicator — institutional desks publish their re-rating overnight, and the tape responds in real time. NOW closed Friday at $90.17 (already +5% off the Wednesday capitulation). A Monday print above $95 confirms the priced-in case advancing. A break below $81.24 reactivates the SaaSpocalypse continuation.
Tuesday April 28 AMC: Alphabet (GOOGL). The platform-scale AI monetization print. Search ad revenue, YouTube ad growth, and Cloud (GCP) AI capex commentary are the three diagnostic lines.
Wednesday April 29 — the day: MSFT, META, AMZN all report after the bell on the same evening — three Mag-7 prints in 90 minutes. Microsoft’s Azure AI growth print is the single most diagnostic line: 35%+ confirms the capex story and extends the AI infrastructure runway; below 30% signals enterprise AI adoption running slower than the semiconductor leadership thesis assumes. Meta on ad-engine AI uplift. Amazon on AWS infrastructure demand. Three-of-three confirming capex acceleration ratifies the within-tech rotation thesis structurally.
Wednesday April 29: FOMC rate decision (3:00 PM ET) + SEP dot-plot release + Powell press conference (3:30 PM ET). Hold expected at 3.75%; the press conference framing is the macro pivot of the quarter.
Thursday, April 30: AAPL AMC + GDP Q1 advance (consensus +2.1%) + Core PCE Mar (consensus 3.1% YoY) + Chicago PMI + initial claims + BoJ rate decision overnight. The single most data-dense day of the macro calendar this quarter — and the reaction to whichever Mag-7 surprised the night before will compound through these prints.
Sunday,y May 4: ServiceNow Analyst Day. Now, an Assist ACV walk-back from $1.5B to $1.0B would confirm that the bridge math the market priced in this week was correct — the AI uplift compensating term collapses, and the SaaSpocalypse case becomes the consensus. Reaffirmation or upward revision would break the bridge math the market just enforced and reset the entire enterprise SaaS multiple back toward pre-Wednesday levels.
Continuing trigger: Rubin Build-Out 100 W/W flipping negative while QQQ continues higher. The within-tech rotation thesis breaks. Capital is leaving the AI infrastructure layer rather than rotating between its sub-sectors. Reduce thematic exposure. The bear case for the entire generation rotation framework would be on the table.
(5) What May Lie Ahead: Macro Setup
Macro setup for next week: the Fed meeting on Wednesday is the structural event of the quarter; the Thursday data cluster (GDP Q1, Core PCE, BoJ overnight, AAPL AMC) is the most data-dense day of the calendar. Money Temperature drifts toward the gross-reduction trigger but has not yet reached it.
Central Bank Pulse
Fed: The Wednesday, April 29 FOMC meeting is the macro event of the quarter — rate decision, SEP dot-plot update, and Powell press conference all in a single afternoon. Market pricing assumes a hold at 3.75%; the question is (a) the dot-plot 2026 median — does the SEP imply two cuts still, or fewer? — and (b) Powell’s press conference framing on disinflation progress. Two crosscurrents: Q1 earnings beat rate (~28% reported, 80%+ beating both EPS and revenue) and resilient retail sales argue against a dovish pivot. Consumer sentiment at 49.8 — the lowest reading on record, below GFC, COVID, and post-Russia-invasion lows — argues for one. Inflation expectations for the year ahead remain elevated at 4.7%. Thursday’s Core PCE +3.1% YoY consensus (a 10bp uptick) and PCE +3.3% YoY (a 50bp uptick) hit the day after the FOMC — making the press-conference framing especially consequential. Fed nominee Kevin Warsh’s confirmation testimony was viewed as more hawkish than expected, which has begun to seep into the rates curve. The DOJ dropped its criminal probe into Powell, removing one source of policy-process uncertainty heading into the meeting.
ECB: The euro area’s dovish path remains intact, though dollar firming this week was less a hawkish-Fed story than a stalled-peace-talks story. The underlying ECB-Fed divergence supporting European equities is unchanged structurally.
BoJ: The Bank of Japan policy meeting is also next week — Thursday, April 30, the morning JST session, with Outlook Report on the same day. Rate-hike expectations have continued to fade in recent weeks. Still, the Outlook Report is the more material item — any hawkish revision would amplify yen-carry-unwind risk on a day already crowded with FOMC follow-through, GDP, Core PCE, and AAPL earnings.
PBoC: China’s Q1 GDP at 5.0% beat expectations. The PBoC’s accommodative stance continues without fresh stimulus. China-listed equities (MCHI) traded mixed on the week, sensitive to the dollar move and to the technology-export-control narrative.
The dollar-yield-oil triangle is the pivot for cross-asset positioning. Last week’s risk-on impulse was anchored on a weakening dollar, falling oil, and stable yields. This week’s sector dispersion was anchored on a firming dollar (UUP +0.44% W/W), rebounding oil (USO +14.10% W/W, a near-complete round-trip of the prior-week's selloff), and modestly higher yields (10Y around 4.30%). Gold also gave back, GLD −2.84% W/W, on the dollar firming. The mechanism is well-understood; the resolution depends on whether the Iran-US peace track resumes after stalling this week.
Macro Commentary
“The path of least resistance for the major US indices is still up. Fade rallies in defensives, lean into the trend in AI infrastructure, and recognize that the rotation thesis depends on the dollar — not on relative-value arguments alone.”
— Michael Hartnett, Bank of America Securities, “The Flow Show,” April 24, 2026
Hartnett’s note this week recommends curve steepeners, China, consumer cyclicals, and frames the AI-infrastructure trade as the dominant US risk-on expression. The AI-infrastructure call aligns directly with the within-tech rotation read — Hartnett identifies the same destination of capital that the Rubin Build-Out captures structurally. The curve-steepener trade presupposes the macro regime supports continued risk appetite, consistent with Money Temperature in the Transition zone rather than Hot. Consumer cyclicals and China are dollar-weakening trades; they sit a step ahead of this week’s price action. They will work when the dollar rolls again. This week’s firming was a pause, not a regime change.
Seasonality
The seasonal calendar (chart shown in Section 3) reinforces the structural read. April 2026 prints +18.54% on QTEC — the best April in the 16-year history. That is one of the largest monthly moves on record, and it happened during a month that the rotation camp had labeled as the moment when “tech leadership is over.” May historically be thin to a coin flip. June is similarly weak. July is the strongest single month of the year for QQQ — average +4.1%, hit rate 94%. August through mid-October enters the seasonal soft window. The structural map for the rest of 2026: a final push higher into summer (sub-5 of larger Wave 3 in the Closelook@Global frame), a sharp setback in late August through mid-October aligning with both seasonality and the historical midterm-year drawdown, and a year-end rally into 2027 from the post-midterm low.
The wave count, the seasonal calendar, and the presidential-cycle pattern all point at the same nine-month roadmap. The current edition is the early stage of that arc.
(6) What May Lie Ahead: Technical Setup
Technical setup: QQQ closed Friday at 663.88, a fresh all-time high. The 4Y log-channel upper band held resistance at 650 last week and was decisively broken this week. The Stochastic remains pinned in the upper range. In a confirmed trend impulse, this is the structural signature, not the warning.
The Technical Reading
QQQ at 663.88 has cleared the 4Y log-channel upper band that capped price last week at ~650. The breakout extended into Friday’s session and held into the close. This is the third impulse of the post-March advance: Wave 1 mid-March breakout from the 520 base, Wave 2 early-April pause at 595-605, Wave 3 the current leg now pressing through the channel ceiling. Invalidation lines: 648 (last week’s prior ATH — a daily close back below would confirm a false breakout); 600 (the 3M consolidation midpoint — close below would argue a deeper retest); 498 (the 4Y log-channel lower band — full structural invalidation).
The setup to watch into next week: a print to 670, followed by rejection back below 658, would activate a 645-650 retest of the breakout level. Sustained price above 670 keeps the structure constructive into the Mag-7 earnings cluster.
Technical Commentary
QQQ at 663.88 has confirmed the breakout above the 4Y channel that capped the price for two weeks. The hold or break at 658-660 on any pullback next week is the primary tactical level — it was upper resistance through Wednesday’s session and now needs to act as support. The reciprocal move requires demand at the prior ceiling, not just the absence of supply. A weekly close above 663.88 — with rising volume — confirms bid conviction into the FOMC. A failure here, with a close below 658, opens a flush toward the 645-650 range where the breakout originated.
The 3M breakout structure at 600 remains the secondary technical floor. A close below that level would end the wave-3 read and argue consolidation into the 580-585 zone. Neither is close at the current 663.88, but a rejection at the upper extension followed by a volume break below 645 would put the 600 test back on the calendar.
Indicator Focus — Why a pinned Stochastic is constructive
The Slow Stochastic (14,3,3) prints in the high range on QQQ weekly, similar to the prior week’s reading. In a confirmed trend impulse, mean-reversion oscillators stay pinned at “overbought” for the entire duration of the move. Every secular uptrend in equity history has looked overbought for most of its duration. The Coppock curve remains rising. ADX remains in a strong-trend reading. The 30W moving average is well below the price and rising. None of these reads are “exit” reads. They are continuation reads in a trending regime.
The diagnostic question is not “is QQQ overbought?” — it is “is this still a trend regime?” The Alignment Check below answers that.
Alignment Check
Dimension Reading Verdict Technical QQQ at 663.88 above 4Y log-channel upper band, fresh ATH, RS to SPY widening, Coppock rising, ADX strong-trend, Stoch pinned overbought ✓ Constructive Macro Money Temperature in Transition zone; Q1 earnings 80%+ beat rate; Fed expected hold; Thursday data day (GDP + Core PCE + BoJ + AAPL) heavy with binary outcomes ⚠ Data-dense ahead Seasonality April 2026 on track for one of the strongest QTEC Aprils on record; May historically thins to a coin flip; July is the destination (94% hit rate, +4.1% avg) ✓ Constructive into early May
═══ Confluent Bullish — Within-Tech Rotation In Progress ═══
All three pillars are constructive. The within-tech rotation thesis is the framework that explains both the software single-name break and the semiconductor extension simultaneously. What to respect: the upper extension of the channel — do not pyramid into 670—what to trust: the trend. The risk-off trigger is Money Temperature 75+ or sector breadth dropping below 50% — neither is on this week’s card.
(7) The AI Build-Out Portfolio: What We Did
Coming Soon
Portfolio moves + accountability log launch with this edition’s next cycle. Subscribers get real-time push notifications on every adjustment before it appears here.
Join the waitlist → closelook.net/c-plus
(8) The AI Build-Out Portfolio: What We Plan To Do
Coming Soon
Forward-plan table with trigger conditions launches alongside §7. C+ subscribers see the candidates’ list before execution.
(9) Knowledge Corner
This Week’s Topic: Why Functional Indices
The week the Tech sector sent two opposing signals at once — semiconductors at 18-day record streaks, software in active retest — is the week the case for functional indices became impossible to ignore. A reader holding XLK could not see the rotation. The ETF averaged it away.
The Problem with Broad Sector ETFs
XLK contains NVDA (semis vertical) and NOW (software cracked) in the same basket. IGV runs a different basket again. SMH a third. If you hold any of them, you get one number that averages multiple divergent stories. In a coherent regime — broad risk-on or broad risk-off — that averaging is fine. The single number tracks the dominant signal. In a within-sector rotation, the averaging is destructive: it hides exactly the signal that matters.
This week’s data illustrates the cost. IGV is down 24.7% YTD. QTEC is up roughly 30%. Both are “tech” by the conventional taxonomy. A reader with XLK exposure participates in the latter through NVDA’s weight and absorbs none of the IGV pain — fine. A reader thinking about whether the broader software complex has bottomed sees only “Tech is up X%” and misses the dispersion entirely. The broad ETF is a useful vehicle for ownership. It is a poor analytical instrument in transition phases.
The Functional Index Approach
A functional index does not start with the question “what stocks have a SIC code in this category?” It starts with “what value-creation function are we trying to capture?” — and then assembles the constituents that functionally deliver that value, regardless of which sector or thematic ETF labels them.
The Closelook Rubin Build-Out 100 captures the AI hardware supply chain — Memory, Packaging, Power, Cooling, Interconnect, Substrates, Lithography, Foundry, EDA. Six layers, 18 sectors. Each sub-sector answers a specific question about the AI build-out: which layer is leading, which is saturating, and which is the next handoff candidate. A reader following the Rubin sector dispersion sees Power Semis up 13.5% this week and knows that GB200 /Rubin power delivery is the active layer. A reader following XLK sees a single number and learns nothing.
This Week as Empirical Proof
The Rubin Build-Out is up roughly 65% YTD on equal weight, ~90% on momentum weight. The Nasdaq-100 is up roughly 6%. A 10-13× spread. That spread is not a quirk of construction. It is the framework working. The Rubin index isolates the AI infrastructure layer. The Nasdaq dilutes the signal with software, consumer, and services exposure. SMH averages across the entire chip complex, including Hopper-era names (Wafer Processing +3% YTD, EDA +16%) and Rubin Early Ramp names (Interconnects +124%, Memory +107%). Rubin separates them.
In a coherent rally where everything moves together, a broad ETF tracks the average,e and that is sufficient. In a rotation phase — Hopper Sunset to Rubin Early Ramp, broad software to narrow orchestration — the broad ETF systematically mispositions. It carries last cycle’s winners and next cycle’s winners in the same basket, weighted by past market cap rather than future capital flow. The functional index disaggregates exactly this distortion.
Beyond Rubin: Functional Indices as a Framework
The Closelook index family is a study in functional Design.
HALO Growth 100 answers: Which physical-world growth compounders sit in sectors that AI agents do not eat? Twelve sectors, deliberately AI-safe, deliberately not-tech. The complement to Rubin in a barbell.
Euro-AI Sovereign 50 answers: which European names are positioned for the EU’s AI sovereignty pivot — chip independence, sovereign data infrastructure, regulated AI deployment? The answer is geographic and regulatory, not sector-driven. No European ETF captures it.
AW25 Agentic Winners 25 answers: Which 25 names are the cleanest sensors for whether agentic software is gaining traction in enterprise workflows? A tactical sensor array, not a buy-and-hold index. Designed to read the orchestration layer the way Rubin reads the infrastructure layer.
ABR Framework answers: which large-cap names are most exposed to AI disruption on five qualitative axes — Legacy Exposure, Agent Benefit, Upside Unpriced, Downside Unpriced, and Confidence? An analytical scoring system that produces an investable index, not a price-screen output.
Each one is built around a question that broad ETFs cannot answer. Each one is reproducible — the methodology is documented, the constituents are tagged, and the rebalancing is rule-based. None of them is “we picked some stocks we like.” All of them are functional designs that pre-commit to capturing a specific value-creation signal.
The Practical Implication
In transition phases — and the Hopper-to-Rubin handoff is the exemplary case for 2026 — functional indices are not a marginal improvement over broad ETFs. They are the only instruments that show the rotation as it happens rather than after it has finished.
When the next cycle handoff comes — Rubin Sunset to Feynman Dawn — the same framework will surface that one in real time. Broad ETFs will report it after.
(10) Upcoming Transactions: Be Informed
Two ways to stay close to Closelook’s portfolio moves — one free, one premium.
Free Access
Follow our work at no cost.
· LinkedIn Newsletter (Closelook@US Stock Markets)
· closelook.net — Free tier
Subscribe → closelook.net/subscribe
Coming Soon · Paid
Real-time transaction alerts before they happen.
· Push notifications on portfolio moves
· C+ Premium Substack section
Join waitlist → closelook.net/c-plus
(11) What May Go Wrong: Risk & Change Triggers
The falsification test: what would invalidate Sections 3, 5, and 6 — and how we would know.
Macro Change Triggers
Wednesday April 29 — FOMC + SEP + Powell: A hawkish dot-plot revision (median pulled to fewer-than-two cuts in 2026) or a hawkish press-conference framing on disinflation progress would lift yields, firm the dollar further, and pressure the multiple-expansion piece of the AI infrastructure trade. Base case: hold at 3.75%, balanced tone, dot-plot keeping two cuts in play.
Thursday, April 30 — the data day: GDP Q1 advance (consensus +2.1% vs +0.5% prior) + Core PCE Mar (consensus 3.1% YoY, +10bp) + PCE Mar (3.3% YoY, +50bp) + Chicago PMI + Initial Claims + BoJ overnight. Both GDP-hot AND Core-PCE-hot in the same morning would force a complete repricing of the rate path — expect a 10Y bid, a firmer dollar, and a risk-off rotation. Either one cooling alone is the base case.
Mag-7 earnings cluster — the AI capex test: GOOGL Tuesday AMC (Cloud + Search), MSFT/META/AMZN Wednesday AMC (the night of three Mag-7 prints in 90 minutes), AAPL Thursday AMC. Microsoft’s Azure AI growth print is the single most diagnostic line — a print below 30% growth would signal enterprise AI adoption running slower than the semiconductor leadership thesis assumes, and the within-tech rotation thesis would face its first fundamental challenge of the year. A print at 35%+ confirms the capex story and extends the runway.
Software-Specific Triggers — The Open Verdict
NOW closed Friday at $90.17, already +5% off Wednesday’s capitulation. IGV at $85.20 held flat on the week — the structural floor at $74.62 sits over $10 below current. The early reads tilt toward bottom-in. The triggers below are the structural decision points.
· NOW close below $81.24 daily: the bridge math collapses, the SaaSpocalypse thesis returns to base case, IGV tests $74.62 next.
· NOW Monday close above $95: the priced-in case advances strongly; analyst re-ratings are net positive; the bottom is in for this episode.
· IGV close below $74.62: the technical floor of the software complex breaks, the March bottom is invalidated, and the AW25 Agentic Winners setup is suspended pending the Q1 earnings sequence completing.
· WDAY (May 27) or CRM (May 28) cRPO deceleration below 15% CC: a sector-wide fundamental problem, not idiosyncratic to NOW. Would confirm that the bridge math applies broadly across enterprise SaaS, not just to ServiceNow.
· Now Assist ACV walk-back at the ServiceNow Analyst Day on May 4: the AI uplift number that supports the bridge math gets revised down. The compensating term in the equation collapses. Bear case for software confirmed.
· Now Assist ACV reaffirmation or upward revision on May 4: the bridge math the market just enforced is rejected by management. The multiple compression from this week begins to unwind. Bull case for software confirmed.
Technical Invalidation Levels (US Indices)
QQQ: weekly close below 648 confirms a false breakout from the channel ceiling. Below 600 ends the wave-3 read and argues consolidation into the 580-585 zone. Below 498 is the full structural invalidation of the multi-year bull case on QQQ.
Indicator Shift Triggers
· Coppock Curve rolling from rising to topping — would flag larger-degree maturity, consistent with the Section 3 forecast that the larger Wave 3 completes in summer 2026. Watch through May-July.
· ADX falling below 20 from a current strong-trend reading — would signal a transition from trending to ranging, consistent with sub-5 completion and a transition to the larger Wave 4 corrective phase forecast for late August through mid-October.
· Rubin Build-Out W/W flipping negative while QQQ continues higher — the within-tech rotation framework breaks. Capital is leaving the AI infrastructure layer rather than rotating between its sub-sectors. The bear case for the entire generation-rotation thesis would be activated.
The thesis holds until one of these triggers fires. When they do, the next edition updates the read.
(12) Final Words
“The best stock to buy is the one you already own.”
— Peter Lynch
Lynch said it about a different kind of market and a different kind of investor. The application to a within-tech rotation, however, is direct: the reader who held the AI infrastructure thesis through the March consolidation, through the rotation-out-of-tech narrative of Q1, through the SaaS scare of this week — that reader did not need to do anything new. The Rubin Build-Out kept holding the right names through every phase, because it was designed around the function the market was rewarding, not around the labels the market was discussing.
The Litmus Test bent — but did not break. The expected revert-to-mean ran its course off oversold extremes. News triggered the mid-week shock. NOW and IBM each took double-digit single-session damage on beats. And IGV at the index level closed essentially flat on the week. The damage was idiosyncratic. Friday’s NOW +5% recovery and SAP’s positive print already tilt the read toward priced-in rather than continuation lower. Tilts are not verdicts, and the earnings response from Tuesday onward is the resolving signal — but the early evidence is on the side of bottom-in, not new lows.
The verdict on software matters specifically for the AW25 setup. It does not matter for the broader thesis. Every “rotation out of tech” call since October 2022 has been a rotation within tech. Episode 47, by Saturday’s close, is data point number nine.
Sub-5 of larger Wave 3 has weeks to run. The Mag-7 earnings cluster is the next test. The FOMC is the macro pivot. The summer push will resolve into the late-August seasonal weakness, the midterm-year setback, and the post-midterm rally — three frameworks pointing to the same nine-month roadmap.
The within-tech rotation is the framework that explains why the software break did not become a tech break, why the semiconductor extension is not euphoria, and why the Rubin Build-Out has tracked at 13× the broad benchmarks year-to-date, and this week confirmed it again. The work is to stay with what works while it is still working — and to respect the falsification levels when they eventually fire.
Thank you for reading. See you next Saturday.
Annex — Further Reading at Closelook
Suggested navigation across Closelook’s own work.
This Week on the Pulse
· The SaaS Litmus Test: Is The March Bottom Real? — Weekly Signal Special, April 19 (closelook.net/weekly/2026-04-19/)
· Rubin Outruns the Nasdaq 100 — By a Factor of Thirteen — Daily Pulse, April 21 (closelook.net/pulse/2026-04-21/)
· The Great Rotation Out of Tech — Season 9, Episode 47 — Daily Pulse, April 23 (closelook.net/pulse/2026-04-23/)
· It ain’t over until it is over — Daily Pulse on NOW/IBM bridge math, April 24 (closelook.net/pulse/2026-04-24/)
Indices & Frameworks
· Rubin Build-Out 100 — 100 stocks, 18 sectors, AI infrastructure (closelook.net/indices/rubin)
· HALO Growth 100 — Physical-world growth, AI-safe sectors (closelook.net/indices/halo)
· Euro-AI Sovereign 50 — European AI sovereignty play (closelook.net/indices/euro-ai)
· Agentic Winners 25 — Tactical agentic infrastructure index (closelook.net/lab/agentic)
· Money Temperature — 8-instrument macro regime scoring (closelook.net/lab/temperature)
· Cointegration Monitor — Engle-Granger pairs + cascade tracker (closelook.net/lab/cointegration)
· ABR Framework — Agent Beneficiary Ratio for AI disruption (closelook.net/research/abr-framework)
Portfolios
· Global ETFs — ex-US regional + sector ETF portfolio (closelook.net/portfolios/global-etfs)
· AI Build-Out — Rubin-index-aligned single-stock portfolio (closelook.net/portfolios/ai-buildout) (this newsletter)
· Hypergrowth — 16 AI-native names + covered-calls layer (closelook.net/portfolios/hypergrowth)
· Global Tech 50 — Non-US tech, passive long-horizon (closelook.net/portfolios/global-tech-50)
· Derivatives — Covered calls, cash-secured puts overlay (closelook.net/portfolios/derivatives)
Editorial
· Weekly Signal — 9-dimensional market regime scorecard (closelook.net/weekly)
· Daily Pulse — Daily market notes archive (closelook.net/pulse)
· Research Dossiers — Deep-dive reports on themes and names (closelook.net/research)
· Closelook 101 — Concepts, frameworks, methodology (closelook.net/101)
· Glossary — Terms, definitions, cross-references (closelook.net/glossary)
Publishing Schedule
