Written theses, scored in public

Tech Conviction Book

Ten written theses on public technology equities, built to bet on selection over index beta, each backed by a price target and a pre-committed thesis-break condition. This is not a portfolio or a fund: no real money is invested. The theses are scored price-only against the Nasdaq-100, held transparently as a demonstration of process. Where a name is in the Library, it links to its full dashboard.
AuthorNaina Garg
InceptionJune 26, 2026
Theses
Pricesloading…
BenchmarkNasdaq-100 (QQQ)
Return since inception
weight-tilted, price-only
Alpha vs. benchmark
vs Nasdaq-100 (QQQ)
Weighted upside to target
12-month price targets
Theses on watch
breached thesis-break

Track Record

Live

The book's return since the June 26 inception, with holdings marked at each nightly close and the benchmark you choose below quoted live.

Benchmark

Cumulative return since inception: Book vs. benchmark

Allocation & Performance

Exhibit I

How the book's weight sits by sub-sector, and which names have actually driven the return.

Allocation

Weights are conviction-tilted, not equal. Larger positions reflect higher conviction and lower idiosyncratic risk.

Contribution to return

Weight × return, in points of the book's weighted return, sorted by contribution. What is actually driving the number, not just what is up: a small position that doubles moves the book less than a top weight that rises 10%.

Run vs. room: return against upside

Each bubble is a holding: horizontal is return since the June 26 entry (how far it has run), vertical is upside remaining to its 12-month target (how much room is left), area is book weight, colour is sub-sector (matching the allocation chart). Dashed lines mark entry, target and the benchmark's return: names to the right of the benchmark line are beating it, names top-left have the most room, bottom-right have run out. A young book, so treat these positions as an early read, not a settled pattern. Full numbers are in the holdings table.

Earned vs. remaining upside

Left arm: return since the June 26 entry. Right arm: remaining upside to the 12-month target. Two different measures on one shared scale, read against each other, never added: a name that has run leans left, a name with room leans right. Marks are the nightly close.

Construction & Concentration

Risk

How the book is built, and how much of it rides on one theme.

Long-only and unlevered. The ten conviction weights sum to 100% with no cash sleeve, sized by conviction rather than equal-weighted.
50%
In semiconductors
5 of the 10 names
71%
Top 5 weights
MSFT, AVGO, NVDA, ORCL, TSM
94%
AI-infra exposed
every name but PANW, on an AI-buildout thesis
6%
Sole diversifier
PANW, the one non-capex leg
This is a concentrated, single-factor book by design. Half of it is semiconductors, and every position but PANW is held for a thesis that depends on AI-infrastructure spending continuing. The conviction tilt is the point, but it means the names tend to move together: the book carries one dominant risk, not ten independent ones.
Book-level thesis-break

Because the book is really one bet, it carries a break at the book level on top of each position's own rule. Either trigger de-risks the whole book, not a single name:

  1. Fundamental. If the four largest hyperscalers guide AI-infrastructure capex flat-to-down for two consecutive quarters, the factor thesis under ~94% of the book is broken and exposure comes down across the board.
  2. Mechanical. If the book draws down 25% from its peak, net exposure is cut regardless of individual theses, a hard floor beneath the per-name 20% breaks.

The Book

Exhibit II

Every name in the book: its weight, current mark, upside left to target, and thesis status.

Conviction
Click a column header to sort. The trend sparkline traces each position's recent nightly closes; tickers in the Library link to their dashboard. "Status" flags ⚠ Watch on a thesis-break (≥20% drawdown) and Near break as a position approaches it. The written rule shows on hover, and in full under Position Theses.

Position Theses & What Would Change My Mind

Exhibit III

The bull case for each name in a sentence, paired with the evidence that would break it.

MSFT Microsoft
Software & Cloud
High

The one mega-cap platform worth conviction weight. Azure+OpenAI is the only hyperscaler where AI monetization is already in the P&L (Copilot ARPU, Azure consumption) rather than just capex, with no search-disruption tail-risk, and it has lagged YTD. The cheapest way to own the hyperscaler beta the index over-owns elsewhere.

Key catalyst
Azure reacceleration + Copilot seat/ARPU disclosure over the next two prints
Key risk
~$190B+ capex outrunning AI revenue; premium multiple
What would change my mind
Azure growth decelerates below ~30% for two consecutive quarters, or Copilot attach/ARPU stalls.
View MSFT dashboard
NVDA NVIDIA
Semis: AI Compute
High

The single irreplaceable AI-compute standard: CUDA lock-in plus the Rubin ramp. Held as our only GPU name and sized as a deliberate overweight (14% vs the index's ~9%), a conviction bet on the one name we won't fade. Whether that overweight adds or costs alpha is what the record will show; the book's differentiation comes from the off-index names it owns instead (TSM, CRDO, ORCL, PANW), not from underweighting the index leader.

Key catalyst
Vera Rubin / NVL72 ramp; >90% data-center growth into rising hyperscaler capex
Key risk
Hyperscaler capex digestion; custom-ASIC / AMD share gains
What would change my mind
Data-center growth rolls over on capex digestion, or a credible >20% custom-ASIC/AMD share shift breaks gross margin below ~70%.
View NVDA dashboard
AVGO Broadcom
Semis: AI Silicon
High

The only at-scale custom-AI-silicon franchise outside Nvidia: AI revenue inflected to $10.8B in Q2 FY26 (+143% YoY, ~half of total) across a six-customer cohort (Google, Meta, Anthropic, OpenAI + two), with management guiding FY27 AI to >$100B, layered on a high-margin VMware software annuity. The post-Q2 ~26% drawdown de-risked the entry, though the bear case ~$344 now sits ~6% below spot.

Key catalyst
Q3 print (the >$100B FY27 AI checkpoint); naming the 5th/6th custom customers; OpenAI 10-GW shipment milestones
Key risk
Priced for perfection (fell ~14% on a Q2 beat); hyperscaler ASIC in-housing; the VMware FY27 renewal wave
What would change my mind
The FY27 AI guide is cut below ~$100B, a lead custom customer pulls in-house, or VMware net retention slips at the renewal wave.
View AVGO dashboard
ORCL Oracle
AI Cloud & Database
High

The cleanest non-mega AI-cloud exposure the index under-weights. OCI's RPO backlog inflection (~$638B, +363% YoY) is re-rating the multiple from legacy-database toward hyperscaler, a re-rating the cap-weighted index does not price. Sized at conviction weight but not co-top, since the thesis depends on that backlog converting cleanly.

Key catalyst
OCI revenue conversion of the RPO backlog + gross-margin stabilization
Key risk
RPO-to-revenue conversion stalls; capex and a July S&P BBB- downgrade pressuring FCF and credit
What would change my mind
OCI gross margins stay depressed or RPO conversion stalls, reversing the re-rating.
View ORCL dashboard
TSM TSMC
Semis: Foundry
High

The cornerstone and the purest embodiment of the mandate: genuinely NOT in QQQ (NYSE ADR), the indispensable ~fairly-valued toll every accelerator vendor pays, ~66% gross margins with 5-10% leading-edge price hikes, and the lowest-correlation quality name available. It captures the entire AI build-out without picking a winner.

Key catalyst
N2 / A16 ramp + leading-edge price hikes; FY26 guide 'above 30%'
Key risk
The un-hedgeable Taiwan single-point-of-failure; rising customer concentration
What would change my mind
A Taiwan-Strait shock, or an AI-capex step-down that idles N3/N2 capacity and breaks leading-edge pricing power.
View TSM dashboard
PANW Palo Alto Networks
Cybersecurity
Medium

The book's sole NON-CAPEX leg, the vector that keeps this from being a pure AI-capex proxy. Platformization (one vendor consolidating network + cloud + SOC) plus AI-SOC drives NGS ARR ~60% with RPO +36%, a recession-resilient, index-under-represented grower, chosen over Hold-rated CRWD's premium multiple. Now trading around our re-rated $365 target (up from $310 on the Aug-14 dashboard refresh, after a run to ~$384), so it stays a 6% Medium-conviction diversifier hold: kept for the non-capex ballast it provides, not for target upside.

Key catalyst
Platformization deal velocity; NGS ARR/RPO acceleration; AI-SOC attach
Key risk
Platformization incentives compressing billings/margins
What would change my mind
NGS ARR growth decelerates below ~40%, or the consolidation deal velocity stalls.
View PANW dashboard
AMZN Amazon
E-commerce & Cloud
Medium

The second platform, held ONLY for its distinct driver: AWS reaccelerated to ~37% (fastest in 18 quarters) plus Trainium custom silicon, and a retail-margin inflection adds a non-AI earnings vector the rest of the book lacks. A real second theme, not a third hyperscaler clone.

Key catalyst
AWS growth sustaining/accelerating + retail operating-margin expansion
Key risk
~$220B capex crushing FCF with no margin payoff
What would change my mind
AWS growth slips back below ~20%, or retail-margin gains reverse with capex still rising.
View AMZN dashboard
CRDO Credo Technology
Semis: AI Connectivity
Speculative

The purest off-benchmark small-cap the Nasdaq-100 entirely misses: active electrical cables (AECs), the under-owned connectivity bottleneck of scale-up AI clusters. Revenue tripled past $1.3B with >80% FY27 growth guided and optical scaling past $600M, on 5 hyperscaler customers. Sized small for the customer-concentration risk.

Key catalyst
AEC design-win expansion to more hyperscalers; the optical ramp
Key risk
Customer concentration (a top customer dual-sourcing); copper→optical transition
What would change my mind
A top AEC customer dual-sources, or the copper/optical shift compresses the AEC TAM faster than optical ramps.
View CRDO dashboard
CRWV CoreWeave
AI Infra: Neocloud
Speculative

The highest-beta direct lever on AI-infrastructure demand: $104B contracted backlog ($129B with early-Q3 commitments, including a $21B Meta deal), a Platinum-tier neocloud, first to validate Vera Rubin NVL72. Now a Nasdaq-100 member (owned for beta/optionality, not as off-benchmark) and sized down for its ~7x leverage.

Key catalyst
Backlog conversion to revenue; Vera Rubin capacity coming online
Key risk
Counterparty/contract risk; GPU rental pricing; debt-funded capex + financing
What would change my mind
A top counterparty renegotiates or cancels, GPU rental pricing softens, or financing tightens against debt-funded capex.
View CRWV dashboard
MU Micron
Semis: Memory
Speculative

The highest-beta way to own the HBM / AI-memory super-cycle: Q3 FY26 printed a record $41.5B revenue at ~85% gross margins with ~$100B of contracted backlog (RPO). Deeply cyclical and owned through the peak: sized small, ride it, don't marry it. A through-cycle DCF (~$575) sits far below the multiples-anchored target, the honest tension a peak-cycle Buy carries.

Key catalyst
HBM4 ramp + pricing; the Q4 ~$50B guide converting; SCA coverage rising above half of revenue
Key risk
Memory cyclicality at a likely cycle peak: HBM ASP/share fade into 2H FY27; a classic boom-bust de-rate
What would change my mind
HBM pricing rolls over or inventory builds: exit on the first hard sign of the cycle turning.
View MU dashboard

Decision Log

Audit trail

A dated, public record of every move in the book, so the process can be checked, not just the result.

How the view has changed

What I got wrong

Publishing theses with dates means being caught out in public. Here is every correction, misjudgment, and change of mind since the June-26 inception, in plain terms: what I thought, what changed, and where I was wrong. Three of these are errors I made and caught myself.

  • Error, self-caughtAVGO: I inflated the target with a share-count error. I divided the sum-of-the-parts by 4.72B shares instead of the 4.94B diluted count. A review caught it; I cut the target ~5% to $461, and the bear case now sits below spot rather than on top of it.
  • Error, self-caughtMU: I published a model that couldn't be true. The DCF's FY27 EBITDA came out below the net income my own EPS implied, which is impossible, and the forward numbers were stale against consensus. I rebuilt it from Micron's real ~80% peak margin and re-rated to $1,400. I should have caught the inconsistency before it went out.
  • Error, self-caughtINTC: I doubted a quote that was right. I flagged Intel's live price as a bad feed. It wasn't: Intel had genuinely re-rated in 2026 (up ~144% year-to-date, and near a 4x peak when I doubted the quote), so the price was real and only my snapshot was stale. The data was correct, and I was wrong to distrust it.
  • MisjudgedPANW: I set the target too low. It ran +117% year-to-date, straight through my $310 target. I under-called the upside I was actually playing for, so the edge is gone. Re-rated Buy → Hold and trimmed it 11% → 6%. Right on direction, wrong on magnitude.
  • MisjudgedORCL: I under-weighted the credit risk. I built the OCI re-rating thesis without pricing the balance sheet hard enough; the July S&P BBB- downgrade forced a target cut from $235 to $190. The thesis holds, my risk-weighting didn't.
  • Changed my mindv1 re-cut: I dropped four names nine days in. The original book held GOOGL, CRWD, ASML and NOW. On a second look none carried enough conviction upside to keep a slot, so I re-cut around higher-conviction, more off-benchmark names.

What isn't here yet: a thesis the market has proven wrong. The book is about two months old, no position has hit its thesis-break, and the sharpest move against me, MU down ~15% from its inception mark, strengthened the case rather than breaking it. Every name carries a written break level and the book a −25% drawdown trigger; when one fires, the re-cut and exactly what I misjudged land here in the same plain terms.

Aug 22, 2026
CorrectionAVGO target corrected $483 → $461 (arithmetic fix, not a re-rate). A review found the SOTP per-share bridge dividing the equity value by the wrong share count. The build now uses Broadcom's actual Q2 FY26 figures: 4.94B non-GAAP diluted shares as the per-share divisor (was 4.72B) and 4.76B common shares outstanding for the $1.74T market cap. The thesis, the segment multiples and the >$100B FY27 AI anchor are unchanged; only the divisor moved, so the base target eases to $461 (+26% from spot), the bull to $535 and the bear to $344. The bear now sits ~6% below spot rather than on top of it, so the downside cushion is thinner than the original framing implied. The book's canonical AVGO target syncs to $461; weight unchanged at 15%.
Aug 22, 2026
Re-rateMU dashboard re-rated Buy $1,275 → $1,400. A verification refresh re-based Micron's stale forward model to current Street reality: price $878 → $967, FY27E consensus ~$158 (we anchor the base below it at $135), and the DCF revenue/margin path lifted to the real super-cycle level (FY27 $135B / 62% → $239B / 80% EBITDA margin), which also fixed a prior inconsistency where the DCF's EBITDA sat below the net income the EPS anchor implied. 12-mo PT $1,275 → $1,400 (10.4x FY27E $135); bull $1,680 → $1,850; bear $620 → $720. The book's canonical MU target syncs to $1,400; the 4% Speculative weight is unchanged.
Aug 20, 2026
Re-rateFour dashboard targets synced. A sweep caught four positions whose dashboards had re-rated since the last sync, and the book's canonical targets were brought current: MSFT $520 → $560 (raised after the Q4 FY26 print), NVDA $215 → $250 (raised as fundamentals strengthened), AMZN $267 → $310 (re-rated on the Q2 AWS reacceleration), and ORCL $235 → $190 (cut on the July S&P BBB- downgrade and the FCF/credit story). Targets are the canonical dashboard figure, so the weighted upside recomputes off the new marks; no position weights changed.
Aug 17, 2026
Re-ratePANW dashboard re-rated Buy $310 → Hold $365. An Aug-14 data refresh caught Palo Alto up ~117% YTD to ~$384, through the old $310 target and even the prior $370 bull case, at ~102x forward EPS with the Street consensus (~$347) below spot. The dashboard moves to Hold at a $365 fair-value target. The book's PANW target syncs to that canonical $365 (from $310); PANW still trades above it, so the position is unchanged: a 6% Medium-conviction ballast hold, kept for its non-capex diversification rather than target upside.
Aug 16, 2026
TrimPANW trimmed 11% → 6%, re-rated High → Medium. Palo Alto has run through our $310 target, so the target upside that justified a top-6 High-conviction weight is gone. It stays as the book's sole non-capex diversifier, held for ballast rather than upside, but at a smaller Medium weight. The freed 5% went to two High-conviction names that still carry real upside to target: MSFT (+3 → 18%) and AVGO (+2 → 15%). The book stays long-only and ~100% invested.
Jun 28, 2026
AddedAVGO + MU added (→ 10 positions). With their dashboards fully rebuilt and 12-month targets re-verified, the two parked Buys join the book. AVGO $483 (+32%), the only at-scale custom-AI-silicon franchise outside Nvidia, FY27 AI guided >$100B, enters at a High-conviction 13%; and MU $1,275 (+13%), the HBM / AI-memory super-cycle owned through the peak, enters as a small 4% Speculative sleeve. Book re-weighted across 10 names (MSFT 15% → MU 4%).
Jun 28, 2026
Held outMRVL and INTC refreshed but kept out. Both dashboards were rebuilt and re-verified, but each has re-rated so far that it is now fairly-to-fully valued: MRVL $255 (Hold) sits just below spot after a +231% YTD run, and INTC $135 (Hold) is up ~3.5x in 2026 on the foundry turnaround with the Street average (~$96) now below the price. Held for the watchlist, not conviction. Correction: the earlier note that INTC's live feed was "returning a bad value" was wrong. The quote (~$128) was real; Intel had genuinely ~3.5x'd, and our dashboard snapshot was simply stale.
Jun 26, 2026
Re-cutBook re-cut for differentiation. Rebuilt from 10 largely-QQQ-overlapping names into a conviction-tilted, deliberately off-benchmark book built to bet on selection over index beta. It opened with 8 positions while four differentiated names (AVGO, MRVL, MU, INTC) were parked pending a dashboard data refresh. Only one GPU name (NVDA, held as a deliberate overweight, the one index leader we back with conviction) and two platforms (MSFT + AMZN, the latter solely for its AWS/Trainium driver). The off-benchmark engine: TSM (NYSE ADR, not in QQQ), the connectivity small/mid-cap the index ignores (CRDO), the AI-cloud re-rater (ORCL), the non-capex security platform (PANW), plus a neocloud beta lever (CRWV). Benchmark unchanged: Nasdaq-100 (QQQ).
Jun 26, 2026
ProcessThesis-break levels are armed for every holding. A position auto-flags to "Watch" on a ≥20% drawdown from entry; its written break condition is monitored and logged rather than auto-detected. Each 12-month target is the canonical figure from the name's dashboard; entries are the live quote at entry, so a fresh position opens at ~0%.
Jun 26, 2026
Priorv1 history. The original 10-name book (NVDA / MSFT / GOOGL / AMZN / AVGO / CRWD / ASML / NOW / MELI / MU, opened June 17) was a largely index-overlapping construction. The MELI→TSM swap (June 26) and this re-cut moved it toward a genuine stock-picker's book; GOOGL, CRWD, ASML and NOW were dropped as their dashboards no longer implied conviction upside.
Every entry, trim, exit, and thesis-break is logged with a timestamp, the "in public" part of scoring theses in public. A position that breaches its thesis-break is logged here automatically. This log is event-driven: a quiet stretch means no position changed, not that monitoring stopped.

Methodology & Disclosure

Not advice. A written, publicly scored expression of conviction: not a portfolio or a fund, no capital, not investment advice.

Construction. Names are chosen off-benchmark on a thesis I can defend, sized by conviction, each with a written break condition, logged and not rationalized when it triggers.

The record. Only a couple of months old, far too short to separate skill from luck: read it as a live process log, not proof of a repeatable edge.

Returns & prices. Price-only from a June 26, 2026 inception (the book starts at 0.0%; the earlier nine-day v1 is logged in the Decision Log and excluded here). Each target is the canonical 12-month figure from its dashboard. Holdings are marked at the same nightly close Open Coverage shows (from coverage-data.js), so the two pages always agree; only the benchmark is quoted live, for the alpha.

Sources. Valuation context from public filings as of late June 2026; the research and this self-updating book were built with an AI workflow.

Part of the Deep Tech Equity Research Library. Browse the 39 company dashboards →