SK Hynix delivered the strongest single-quarter print in memory chip history: 72% operating margin, revenue +198% YoY, and a structural HBM shortage forecast through 2030. The semiconductor index logged its 18th consecutive up day — a record streak. SMH closed at $506.58, up 9.3% in five sessions. The AI thesis is no longer "directional" — it's structural through the decade. Real yields eased 4bp to 1.89% (away from the GLD trim trigger). The ceasefire was extended indefinitely on April 21 with active diplomacy in Pakistan this weekend. Two binary catalysts ahead this week: COPOM Monday, Fed Wednesday.
The April 23 print was beyond a beat — it was a regime input. Revenue 52.6T won (+198% YoY), operating profit 37.6T won (+405% YoY), 72% operating margin. Net cash position 35T won. But the headline isn't the print — it's Chairman Chey Tae-won confirming HBM shortage persists to 2030 with 20%+ capacity gap, and the company stating customers are "prioritizing procurement over price." HBM prices +50% YoY. This is multi-year regime input, not a single-quarter event. Combined with TSMC's Q1 beat April 16, both pillars of the AI hardware thesis are confirmed for at least the next four years. EWY conviction stays at 9/10 (held at 12%); SMH valuation gate remains lifted (P/E 22x). The thesis is no longer "AI is real" — it's "AI is supply-constrained through 2030."
All 11 mandatory BRICS searches completed. Real yield checked [v4-1]. Valuation gate lifted [v4-4]. USD/JPY zone [v4-2]. AMLP exit confirmed [v4-3] — Henry Hub at $2.52, 18-month low. Copper checked [v4-5]. VGK held [v4-6]. Options expiry rule applied [v4-7]. Protocol v4 — no step skipped.
After a vertical week — semis +18 sessions, S&P/Nasdaq ATH, SMH +9% — the prudent move is not to chase. Two binary catalysts in five days: COPOM Monday, Fed Wednesday. ITA trimmed 4% → 3% as the indefinite ceasefire reduces tactical war premium. BIL raised 15% → 16% as buffer. AI sleeve held at 32% — conviction is structural but momentum is parabolic. EWZ held at 13% pending COPOM. SK Hynix-validated EWY held at 12%.
| Ticker | Sleeve | Weight | Conviction | Rationale |
|---|---|---|---|---|
| EWY | AI / Tech Primary | 9/10 |
SK Hynix Q1: 72% margin, +198% YoY, HBM to 2030. KOSPI 6,475 (intraday ATH 6,557 April 23). EWY $154.57 (+1.5% week). Q1 GDP +1.7% — fastest in 5+ years. Held at 12%; structural thesis confirmed but no chase at ATH. | |
| SMH | AI / Tech Primary | 10/10 |
$506.58 (+9.3% week). Valuation gate remains lifted (P/E 22x). 18 consecutive up sessions = record streak. Conviction MAX but momentum parabolic. Held at 12% — no chase at fresh ATH despite full-conviction thesis. | |
| TQQQ | AI / Tech Primary | 7/10 |
S&P 7,165 ATH. Nasdaq 24,837 ATH (+1.5% week). VIX 18.71 (12 days <20). Leveraged tech captures broad market upside. Cap at 8% — leverage risk through Fed FOMC Wednesday. Held. | |
| GLD | USD Debasement | 8/10 |
Spot $4,710 (−3% week). Real yield eased to 1.89% — moved further from 2.0% trim trigger. Structural CB bid intact (755–800t 2026 forecast). Held at 12% — real yield improvement offsets price weakness. Fed Wednesday is the binary. | |
| GDX | USD Debasement | 6/10 |
ZAR strengthening trend pre-week (paused with gold pullback). 1.5–2× beta to spot gold means deeper drawdown than GLD this week. Held — structural thesis intact at 1.89% real yield. | |
| EWZ | BRICS / EM | 7/10 ↓ |
EWZ $39.94 (−3% week). Ibovespa MACD bearish crossover. BRL R$4.98 — rejected R$5.02 retest, still below R$5.00. IPCA 2026 forecast lifted to 4.80% (above target ceiling). COPOM Monday: 25bp cut expected. Held — conviction notched down to 7. | |
| FXI | BRICS / China | 4/10 |
$36.86 (+1% day). PBoC LPR unchanged 11 months. Chinese smelters record refined copper March (1.33Mt). Yangshan copper premium $68/ton. Domestic demand stable. Held at 4%. | |
| VGK | Deglobal. | 7/10 |
Germany €127B active [v4-6]. EUR/USD ~1.18. ECB pricing 2+ hikes. VGK +33% YoY. Oil $105+ creates near-term margin headwind but structural rearmament + fiscal expansion intact. Held at 8%. | |
| ITA | Defense (residual) | 5/10 ↓ |
Trimmed from 4%. Ceasefire extended indefinitely April 21. Active diplomacy this weekend. Tactical Iran premium reduced. NATO/EU structural rearmament thesis retained at 3%. Full exit if Saturday talks succeed and oil breaks below $90. | |
| IBIT | Alt Monetary | 6/10 |
Bitcoin $77,975. IBIT $43.92. $167.5M inflows on April 24 — renewed conviction. AUM $63.1B. USD debasement amplifier when real yields ease. Held at 5%. | |
| BIL | Liquidity | N/A |
Raised from 15%. +1% from ITA trim. Two binary catalysts this week: COPOM Monday April 27, Fed FOMC Wednesday April 29. Saturday Witkoff/Kushner Iran talks could move Monday open. Buffer against parabolic SOX mean-reversion. ~4% yield while waiting. | |
| [OPT] GLD May ATM |
6/10 ↓ |
Gold sold off to $4,710 — call now ~3% OTM (was ITM April 17). Real yield improved to 1.89% (positive for gold). Hold through Fed Wednesday. Exit if GLD < $4,600 or real yield breaks 2.0%. Conviction notched down to 6. | ||
| Total | 100% | AI/Tech 32% · Debasement 17% · BRICS/EM 17% · DG 8% · Defense 3% · Alt Mon 5% · Liquidity 16% · Options 2% | ||
Changes from April 20
SMH is up 9.3% in five sessions. The SOX index has logged 18 consecutive up days — a record streak. The valuation gate is lifted. Conviction is at maximum 10/10. And yet adding here is wrong. The structural thesis is bulletproof through 2030, but Mondays after 18-day vertical moves are statistically the worst entry points. Per protocol Phase 3 contrarian check: "> 60% in one regime?" — AI/Tech sleeve at 32% is fine, but adding more SMH here would be chasing momentum at the exact moment mean reversion is most likely. Wait for a 3–5% pullback before sizing up. The conviction is for the regime, not the entry.
Henry Hub $2.52/MMBtu on April 24 — lowest since October 2024. Down 13.4% over the past month. The protocol $3.50 unconditional exit threshold remains broken by a wide margin. Oil at $105+ has not transmitted to US domestic gas because storage is 8% above seasonal norms and producers like EQT are curbing output. The AI gas demand thesis simply has not materialized at the Henry Hub spot level. Reinitiation requires Henry Hub sustained above $5.00 for two weeks — currently nowhere in sight.
Nifty 23,898 (April 24) — down from 24,353 (April 17), and well below the 52-week high of 26,373 (January 5). The Nifty is now further from its 3-month high than last session. Worse, INR weakened to 94.25 — 5 consecutive days of losses, well above the 90 EPI threshold. Both conditions of the EPI initiation rule are failing in the wrong direction. India is explicitly excluded for a third consecutive session. Watch for an oil-driven reversal if Saturday talks succeed.
Brent +14% week to $105.33. WTI $94.40. Tempting, but the position is a binary on Saturday's Witkoff/Kushner talks. A successful diplomatic breakthrough could send oil back to $80 within a week. The 3% ITA residual captures the geopolitical premium without the binary commodity risk. If Hormuz reopens, oil crashes — any long oil position taken at $105 would be crushed.
BRL closed Friday at R$4.98 — below R$5.00 but rejecting from R$5.02 retest. Not yet at the R$4.90 mandatory ADD signal. Meanwhile, Boletim Focus raised 2026 IPCA forecast to 4.80% (above the 4.50% target ceiling) and Selic end-2026 to 13.00% (from 12.50%) — a hawkish revision. Ibovespa MACD turned bearish for the first time since the rally began. EWZ at 13% is already at the upper end of my target range; adding before COPOM (Monday) or before R$4.90 breaks would be premature. Trim signal at R$5.50, exit at R$5.80, ADD trigger only at R$4.90.
Saturday April 25: Witkoff/Kushner Pakistan talks (could move Monday open). Monday April 27: Brazil COPOM (25bp cut to 14.50% expected). Wednesday April 29: Fed FOMC (no action expected, 39% Dec cut probability). Real yield 1.89% — 11bp from GLD trim trigger. 50% trim on first break signal; full exit on multi-session confirmation.
Real yield breaks above 2.0% and holds 5 sessions → trim GLD 50% immediately. Currently 1.89% — improved 4bp this week.
Breaks above 2.5% → full GLD exit + GLD call exit same day.
Hawkish Warsh signal at FOMC = key risk factor.
No action expected. 39% probability of December cut (up from 23%). Watch for Warsh nomination signal.
Hawkish surprise → exit GLD call same day, trim GLD to 9%, raise BIL to 18%.
Dovish surprise → roll GLD call higher, deploy BIL into SMH/EWY toward 14% each.
Base case: 25bp cut to 14.50%. IPCA 2026 forecast lifted to 4.80% (above target).
Hawkish hold or 25bp + hawkish guidance → BRL strengthens, EWZ holds 13%.
50bp cut surprise (low probability) → BRL weakens, trim EWZ 50% same session.
SOX has logged 18 consecutive up sessions — a record streak. Statistical mean reversion overdue.
SMH closes below $485 for 2 sessions → reduce SMH 1–2% on technical breakdown.
SMH closes below $465 → trim SMH to 9%; momentum thesis broken.
BRL R$4.98 — close to R$5.00 pivot.
BRL weakens beyond R$5.50 → trim EWZ 50%.
BRL weakens beyond R$5.80 → exit EWZ entirely.
BRL breaks below R$4.90 → mandatory ADD EWZ (Brazil Rule).
USD/JPY ~158 — still in Zone 3, BoJ intervention at 160. Range-bound.
Breaks below 148 → raise BIL 5%, cut FXI.
Breaks below 140 → carry unwind extreme; raise BIL 10%, exit FXI.
Saturday talks succeed → oil falls below $90, ITA exit, BIL deploy aggressively. SMH could test $530+.
Brent below $90 sustained → AI sleeve raise toward 35% (SMH → 14%, EWY → 14%).
This is the weekend's largest single-event amplifier.
Q1 print delivered structural validation: 72% margin, +198% YoY, HBM gap to 2030.
Magnificent 7 earnings in coming weeks could amplify if hyperscaler capex guides up.
EWY maximum target 14%; structural thesis intact for years.
Real yields drop below 1.7% → raise GLD toward 14%.
PBoC buying confirmed above 900t annualized → raise GLD to 15%.
Dovish FOMC surprise → roll GLD call strike higher to lock in.
BRL breaks below R$4.90 → mandatory additional EWZ. Brazil Rule requires adding.
COPOM dovish guidance Monday + commodity tailwind → EWZ to 14%.
Currently R$4.98 — one strong session away from add signal.
Currently failing both conditions: Nifty 23,898 (below 3-month high), INR 94.25 (above 90).
Witkoff/Kushner success could reverse INR weakness instantly.
EPI initiation requires Nifty above 3-month high AND INR below 90 simultaneously — monitor for rapid post-Iran reversal.
SMH closes 3–5% below current $506 on healthy mean reversion → add SMH toward 14%.
Conviction is at structural maximum after SK Hynix print; the right move is to add on weakness, not at ATH.
Combine with EWY to lift AI sleeve to 35% on a ~$485 SMH retest.
DXY 98.52. EUR/USD ~1.18. USD/JPY ~158 (Zone 3). BRL R$4.98 — rejected R$5.02 retest. INR 94.25 (5-day weakness). Per protocol cross-BRICS rule: BRL flat + INR weakening = "partial confirmation only — maintain current allocations." 11 mandatory searches complete.
Weakest Link
Holding SMH at 12% after an 18-session vertical move at fresh 52-week high is the single largest risk in the portfolio. The structural thesis is bulletproof but the price action is parabolic. If a Saturday talk failure plus a hawkish Warsh signal Wednesday combines into a single risk-off event, SMH could drop 8–10% in three sessions and the portfolio loses ~1% of NAV from this position alone. The "no chase but no trim" decision splits the difference but won't be perfect either way.
Most Underpriced Risk
The GLD May call is now OTM. The protocol invalidation is GLD < $4,600 (currently $4,710 — only $110 away, ~2.3%). A hawkish Warsh-driven FOMC could push real yields above 2.0% and gold below $4,600 in a single session. The "win in 3 of 4 scenarios" framing assumes the bad scenario is bounded — but a true hawkish surprise from a new Fed Chair nominee is fatter-tailed than the framing implies. The 2% premium is at-risk.
Narrative Not Fully Considered
What if the SK Hynix print is being read backwards? +198% YoY revenue growth and 72% operating margins are ATH-level metrics that historically mark cycle peaks, not bases. The "shortage to 2030" claim is a CEO statement, not market data. The semiconductor index up 18 days straight could be the late euphoric blow-off rather than the start of a new leg. If memory pricing peaks and HBM4 capacity ramps in 2027 (per SK Hynix's own roadmap), the 2030 shortage thesis could compress. I'm assuming structural, but it could be cyclical at peak.
BRICS-Specific Risk
EWZ at 13% with BRL stalled at R$4.98 and Ibovespa MACD turning bearish is a structural blind spot if COPOM delivers a hawkish hold or 25bp + hawkish guidance. Brazil's IPCA forecast is now above the 4.50% target ceiling — the BCB has less room to ease than the April rally priced in. If Selic stays at 14.75% Monday or guidance is "one and done", BRL could weaken back through R$5.00 and Ibovespa could test 187,000 (200-day SMA). EWZ -5% is the realistic downside in that scenario.
Published when the macro changes — not on a schedule. Fed FOMC Wednesday April 29 will likely trigger the next session within 24 hours.