AMUU

Current Thesis

AMUU is a 2x-daily-reset leveraged wrapper on AMD — there is no independent thesis, every PnL driver is AMD's narrative amplified. Right now AMD is on an 11–12 session win streak (longest since 2005), printing ATH ~$281 into a Q1 2026 earnings print on 2026-04-30. The narrative leg is the MI350/MI355X ramp plus the MI400 H2'26 launch, OpenAI + Meta anchor wins, and the Q4'25 milestone where GPU revenue first eclipsed CPU revenue. For us the trade is simple: we would be chasing a parabolic, retail-saturated leveraged vehicle 10 trading days from a binary print. That is the textbook beginner trap, not an entry.

Bull Case

• AMD Q1'26 guide: ~$9.8B revenue (+32% YoY), ~55% non-GAAP GM, with ~$100M MI308 China re-start baked in. A beat + raise Q2 (>$10.5B) confirms the accelerator ramp and sends AMUU +15–25% on a single session (2x compounding on a +8–12% AMD gap).
• MI400 (CDNA-Next, 432GB HBM4, 19.6TB/s, 40 PF FP4) on track for H2 2026 — this is the Nvidia-parity chip on paper. If the Q1 call confirms MI400 tape-out / customer LOIs, the narrative transitions from "AMD is a credible #2" to "AMD owns the inference TAM."
• Sell-side catch-up incomplete: avg PT $289 (vs spot ~$278), Bernstein $365, 37/37 buy ratings. Room for upgrades to push towards $320–$350 post-print.
• Meta partnership (per Rasgon) is structurally underappreciated — multi-year accelerator commitment, not a one-quarter headline. This is the kind of datapoint that retail is still pricing in.
• AMD now trades above 20-EMA ($235) by ~18% — AMUU, on 2x daily reset, is ~30–40% above its own implied 20-EMA. Momentum of this shape historically persists into the next quarter when earnings confirm the ramp.

Bear Case

• The wrapper is the trade's biggest flaw. 2x daily-reset ETFs have volatility decay: a +5% / –5% round trip in AMD loses ~0.5% per cycle in AMUU. Holding periods >1 week in choppy tape are a guaranteed drag even if AMD ends flat. Expense ratio 0.97–1.06% piled on top.
• 11-day win streak → RSI almost certainly >75 on both AMD and AMUU. That is an archetype-6 momentum blowoff signal by our own playbook (Core Principle 5c: trim on RSI>75 for squeeze names).
• Q1 print on 2026-04-30 is a binary. Sequential revenue is down 5% per AMD's own guide — if whisper numbers have drifted higher, a simple "in-line" becomes a sell-the-news. 2x leverage makes every disappointment scenario catastrophic (a –10% AMD gap = –20% AMUU).
• MI308 China is a political football — another export control tweak zeros out ~$700M in FY'26 revenue and the stock loses the China optionality leg.
• Competition risk: Nvidia GB300 rack-scale ships in volume in Q2; hyperscaler custom silicon (TPU v7, MTIA, Trainium3) continues to cannibalize the merchant AI-accelerator TAM. Any benchmark where MI355X underperforms GB200 cost/token breaks the parity narrative instantly.
• Leveraged single-stock ETFs are the definition of peak retail vehicle — if AMUU volumes are spiking 3–14d, that is the saturation tell, not the acceleration tell.

Setup & Price Structure

• AMUU ~$78 (2026-04-17). AMD spot ~$278 vs 20-EMA $235 → 18% stretched on the underlying, implying ~30–40% stretched on the 2x wrapper.
• 11-day consecutive green close on AMD = the longest streak since 2005. By base-rate, day 12+ has a negative expected return over 5 sessions.
• Clean bullish structure: no recent lower-low on AMD, higher-high daily, parabolic curve. But curve is now vertical — the kind of tape where the first red candle starts a 15–20% AMD drawdown, which would be a 30–40% AMUU drawdown.
• No constructive pullback to buy. The only clean re-entry on AMUU is a post-earnings reset: either (a) gap up, hold above prior high 3 sessions, re-enter on retest; or (b) gap down, flush, AMD re-claims 20-EMA, re-enter at breakout retest.
• Do not chase vertical. Do not buy 2x leverage into a 10-day earnings window on a 2x-stretched-above-20-EMA name. All four beginner-trap conditions from the playbook trigger at once.

Catalyst Calendar (next 30 days)

• 2026-04-30 — AMD Q1'26 earnings (AMC). Consensus ~$9.8B rev, ~$0.94 EPS, 55% GM. Guide for Q2 is the real tell — looking for >$10.5B to confirm the accelerator step-up.
• ~2026-04-22 to 2026-04-30 — AMD pre-announcement / analyst-day headline risk window.
• 2026-04-30 (post-close) → 2026-05-01 open — AMUU's defining 24h. Implied move on AMD is likely ±8–10%; AMUU move is ±16–20%.
• 2026-05-07 (est.) — NVDA earnings read-through (NVDA reports late May but pre-announcement commentary matters). NVDA beat + raise is good for AMD sentiment; NVDA guide warning is bad for AMUU despite being AMD-adjacent.
• 2026-05 (TBC) — Computex / AMD developer event window historically used for MI400 teasers.

What Would Change Our Mind

• Invalidation triggers (specific, observable):
• AMD weekly a daily close below the thesis-invalidation level (20-EMA) → trend break, AMUU structurally broken at ~$60 → hard exit.
• AMD Q1 revenue < $9.5B (low end of guide) OR Q2 guide < $10.2B → thesis breaks on Apr-30 print, AMUU –15–25% gap, do not average down.
• AMD-specific insider selling cluster (>$50M) in next 14d → smart money taking the other side of the run.
• MI355X benchmark leak showing <60% of GB200 cost/token → parity narrative dead.
• Re-engagement triggers (what gets us long after a reset):
• Clean post-earnings gap-and-go (AMD +5% minimum, holds 3 sessions above the gap) → scale in to AMUU 1–2% sizing.
• AMD pullback to 20-EMA ($235) WITHOUT broken structure, with MI400 catalyst still 60–90d out → probe AMUU at the 20-EMA retest.
• Up-size to HIGH conviction only if: (a) we're past the Apr-30 earnings event, (b) AMD is in a fresh basing-breakout above pre-print high, (c) analyst revisions upward-clustered, (d) AMUU is not more than 20% above its own 20-EMA.

Correlation Notes

• 100% beta to AMD (by construction). AMUU is not a diversifying holding — any existing AMD / NVDL / SOXL / USD semi-exposure stacks directly.
• Highly correlated to NVDA, AVGO, MRVL, SMCI, and the AI-accelerator complex. Any holdings in those names reduce AMUU sizing budget.
• Secondary correlation to TSM (AMD's foundry dependency), MU/HBM names (MI355X/MI400 memory bill-of-materials), and ARM (datacenter CPU pairing).
• Macro: hyper-sensitive to 10Y yields (duration proxy on high-multiple growth), USD/CNY (China export regime), and any Fed surprise. 2x leverage amplifies every macro shock.
• Treat AMUU as a pure AMD options-lite substitute: sizing should be set on AMD-equivalent exposure (1% AMUU ≈ 2% AMD notional risk), NOT on account %. Archetype-6 cap of 1%/name is the correct ceiling.