A failed Korea IPO plus a bank stress-testing a CoreWeave default — short the debt-heavy AI data-center stocks

Verdict: the CoreWeave credit story is real, but the trade isn't live — wait for the confirmed breakdown

The idea's bearish thesis has genuine filing-level support: CoreWeave carries $14.7B of long-term debt against $3.3B of equity, free cash flow was negative $7.3B for fiscal 2025, net margin worsened 6.9 points quarter over quarter to negative 35.6% in Q1 (period ended March 31, 2026), and shares outstanding jumped 5.9% to 532 million — the dilution-and-reborrow mechanic the thesis describes, with a major bank publicly modeling a CRWV default per the MarketWatch piece from October 20, 2026. The strongest point against is that the short's second leg undermines it: Nebius swung to an $82.5M net profit on 479% revenue growth with gross margin up 16.4 points to 68.6%, so a pair-trade framing is fighting improving fundamentals on one side, and the strategy's own entry rules never fired across 247 daily bars in the trailing 12 months. This is a rules-not-triggered watch-list setup, not an active signal: CRWV closed at $62.13, just above the $62.00 support, but the trend-strength filter reads 21.6 versus the required 25 and the volume-flow condition is unconfirmed. No robust parameter setup was established because the evaluation could not complete, so the published thresholds are the levels to watch, and a short into double-digit volatility with a 2.7% position-level stop and no live trigger earns a wait. The next CRWV/NBIS quarterly filings and the December 31, 2026 ownership cycle are the dated events that will confirm or break the credit narrative.

Trade now: CRWV breakdown still waiting on its trigger

Nothing is live yet. The idea is a bearish, short-focused setup on CoreWeave (CRWV) with Nebius (NBIS) as the secondary name, but the entry rules have not fired: over the last 12 months of daily bars (247 evaluated), the strategy recorded zero entries. That is a watch-list state, not a trust problem — the rules were run on real bars and simply found no qualifying trigger. The entry stack for CRWV requires a close crossing below the first support level plus confirmation conditions, and the market is close but not there. CRWV closed at $62.13, with the nearest support at $62.00 — only $0.13 away. But the trend-strength filter needs the ADX (14) at or above 25, and it currently reads 21.6, about 3.4 points short. The on-balance-volume condition is also not confirmable right now. NBIS is even further away: last close $171.45, nearest resistance at $168.71, ADX at 18.7 versus the 25 threshold — roughly 6.3 points short. NBIS's RSI of 26 versus CRWV's 18.8 tells you CRWV is already the more stretched name, consistent with the short-into-momentum thesis. Concrete risk framing once an entry does fire: the strategy's own guardrails are a hard stop at a 2.7% loss on the position and a take-profit at a 5.4% gain — an effective reward-to-risk of roughly 2:1 — with a maximum single-position size of 25% of capital and a 45-day time-based exit. In price terms, a CRWV entry near $62.13 implies a stop zone around $60.47 and a target near $65.46, but that only becomes actionable after the trigger conditions align. So "wait" means exactly this: no position until a CRWV daily close breaks below $62.00 while ADX is at or above 25 and the volume-flow condition confirms. If CRWV keeps bleeding (it is 55.6% off its range high and only 2.2% above its range low) without ever printing the confirmed breakdown, the trade never opens — that is the discipline, not a gap. On parameters: the research author requested a bounded local search around the current thresholds, but because the frozen rules could not be fully evaluated, no robust setup was established — so the published thresholds are the ones to watch.

The Leverage Math Is Starting to Scare the Lenders

The idea argues that the most leveraged 'neocloud' names fall hardest when AI spending doubts spread, and CoreWeave's own filings supply the leverage. At fiscal year-end 2025, CRWV carried $14.7B of long-term debt against just $3.3B of equity, with free cash flow of negative $7.3B for the year — a run-rate that has to be funded with new borrowings or new shares. The Q1 2026 update shows the company chose dilution: shares outstanding jumped 5.9% in a single quarter, from 502.4 million to 532 million. That is the exact mechanic the thesis describes — a business model that depends on cheap debt and sustained demand, forced back to capital markets as conditions tighten. The direction of the most recent fundamentals is unambiguously worse. Quarter over quarter into the March 2026 period, gross margin fell 2.1 points to 65.5%, operating margin slipped 1.2 points to negative 6.9%, net margin deteriorated 6.9 points to negative 35.6%, and return on equity worsened 2.0 points to negative 15.5%. Revenue kept growing ($1.6B in Q4 2025 to $2.1B in Q1 2026), but growth funded by widening losses and rising share count is precisely the profile that gets repriced hardest when the demand narrative cracks. The cited news adds a credit dimension. Per the MarketWatch piece from October 20, 2026, JPMorgan ran the numbers on what a CoreWeave default would do to Wall Street — and per Bloomberg's same-day coverage, AI fears 'roar back' heading into tech earnings. The thesis reads that correctly: when a major bank publicly models your default, lenders themselves are pricing real credit risk. A short into continued negative momentum has a live catalyst set. Nebius, the second name in the idea, reinforces the pattern from the ownership side. The September…

Scores

• Conviction score breakdown: 54
• Thesis support: 72
• Trade readiness: 38
• Risk quality: 45
• Trigger proximity: 55
• Fundamentals trend: 60

Watch items

• CRWV — Close vs first support level
• CRWV — ADX (14)
• CRWV — On-balance volume vs price
• CRWV — Net margin (Q1 2026)
• CRWV — Shares outstanding
• NBIS — Insider net open-market activity
• NBIS — ADX (14)
• NBIS — Net income (FY2025)