NVDU

Current Thesis

NVDU is a 2x DAILY-RESET leveraged ETF on NVDA (expense ratio 0.92%) — a pure-momentum vehicle on the dominant AI-chip narrative, NOT a company. The narrative leg we'd buy: NVDA's data-center supercycle is still accelerating on the second derivative of dollars even as % growth decelerates. Q1 FY2027 (printed 2026-05-20) did $81.6B rev (+85% YoY), DC $75.25B (+92% YoY, +21% QoQ), EPS $1.87 vs $1.77 est, 74.9% gross margin, $49B FCF, dividend hiked 25x. Q2 FY2027 guide $91B ±2% — >$4B above consensus and with zero China DC revenue assumed. CFO cites $500B Blackwell+Rubin visibility through CY2026. The catch for a 2x vehicle: NVDA went into the print at ~$224 and now sits ~$215–222 (2026-06-03) — a post-earnings digestion/fade. Chop is poison for a daily-reset 2x ETF. NVDU at $151.09 is ~13% under its 52wk high ($172.89), i.e. consolidating, not breaking out. Full conviction needs NVDA to reclaim the post-print high $232.

Bull Case

• 2026-05-20 Q1 FY2027 blowout: rev $81.6B (+85% YoY) vs the law of large numbers, DC $75.25B (+92% YoY / +21% QoQ), GM 74.9%, record $49B FCF, dividend raised 25x. No deceleration in the print itself.
• 2026-05-20 Q2 guide $91B ±2% — $4B+ above the Street, GM held at 75%. A raised guide above consensus is the textbook ACCELERATING tell; this is confirmation, not a reason to wait.
• $500B Blackwell+Rubin revenue visibility (CFO, CY2025→end-2026) — locked backlog de-risks the next 2 prints.
• Vera Rubin entered full production Q1 2026; partner availability 2H 2026 → fresh product-cycle catalyst into year-end.
• China optionality: guide assumes $0 China DC revenue — any thaw is unbudgeted upside.
• Analyst frame: 37 analysts Buy, consensus NVDA PT ~$298 (≈38% over $215) → at 2x daily in a clean trend that's a ~70%+ NVDU leg. NVDU already ran $77.12→[trade redacted] over 52wks.

Bear Case

• Mechanical decay is the #1 risk — 2x DAILY reset + 0.92% fee means NVDU bleeds in sideways/choppy NVDA tape even if NVDA is flat. This is structural, not narrative. The current post-print consolidation is exactly the regime that grinds a 2x ETF lower.
• "Sell the news" already in motion: NVDA $224 pre-print → ~$215–222 after a beat-and-raise (2026-06-03). Blowout + fade = digestion.
• % growth decelerating: +85% YoY now vs +100%+ a year ago — second-derivative of rates is rolling even as dollars grow.
• Customer concentration: hyperscalers >50% of DC rev ($38B, +12% QoQ). One capex-guide cut from MSFT/GOOGL/AMZN/META cracks the whole theme.
• Custom-silicon erosion: Broadcom/Google TPU/Amazon Trainium + AMD MI400 chipping at merchant-GPU TAM.
• Year-3 mania risk: AI-capex trade is maturing; any digestion phase is amplified 2x and path-dependent on NVDU.

Setup & Price Structure

NVDU $151.09 (2026-06-03), intraday range $149.93–$153.43. 52wk range $77.12–$172.89 → currently ~13% below the high = consolidation, not breakout. Underlying NVDA ~$215–222 (close ~$214.75, intraday to $221.79), faded from the $224 pre-print level, with a post-earnings intraday high of $232.28 acting as the line in the sand. Structure read: NVDA is range-bound digesting a beat-and-raise. For a 2x daily vehicle this is the worst tape (decay with no trend). The clean trigger is NVDA reclaiming $232 on volume — that flips NVDU from MEDIUM probe to HIGH trend-add. Below NVDA ~$195 the post-Q1 base breaks. Do NOT average down a 2x ETF — broken trend = compounding works against you fast.

Catalyst Calendar (next 30 days)

• No NVDA/NVDU-specific binary in the 30-day window — Q1 already printed 2026-05-20. This is a clean ~12-week no-earnings runway, a positive (no binary landmine) and a key reason the trade can be carried.
• ~2026-06-17 (est.) FOMC — rate path moves high-beta/long-duration AI complex; a 2x leveraged vehicle is macro-sensitive here.
• ~2026-08-27 (est.) Q2 FY2027 print — the next true binary; revisit blackout 3 trading days prior.
• 2H 2026 Vera Rubin partner-availability headlines (undated) — narrative refresh, not a scheduled catalyst.

What Would Change Our Mind

• NVDU daily close below $135 (breaks consolidation floor) → exit, no debate.
• NVDA weekly close below $195 (post-Q1 base) → trend broken.
• Any hyperscaler cuts AI-capex guidance (MSFT/GOOGL/AMZN/META) → demand-side crack, theme flips toward SATURATED.
• NVDA Q2 (Aug) DC revenue QoQ decline or guide cut → narrative break, full exit.
• AMD MI400 / custom ASIC wins meaningful share at a top-3 hyperscaler.
• NVDA range-bound >4 weeks → exit NVDU on decay grounds alone even if thesis intact (instrument-specific rule, not a narrative call).

Correlation Notes

NVDU is ~2x daily beta to NVDA — effectively a concentrated single-name factor bet, amplified. High correlation to the AI-chip/compute complex: AVGO, AMD, MU, TSM, SMCI, plus the power/infra second-order names (VRT, ETN) and the hyperscaler demand block (MSFT, GOOGL, META, AMZN). Never pair NVDU with another NVDA-levered or AI-chip long — that's doubling/tripling the same trade and the 2x makes the drawdown brutal. Inverse vehicle: NVDD (1x bear NVDA). Tracks SOX/SMH directionally but carries 2x idiosyncratic single-name + path/decay + swap-counterparty risk on top.

Theme State

ai-chip-infra: ACCELERATING (Q2 guide raised above consensus, $500B visibility) but with MATURING characteristics (year-3, % growth rolling, post-print digestion). Instrument timing right now = consolidation, not a fresh breakout.