1
1Wednesday’s trading session painted a clear picture: investors are punishing AI infrastructure companies that borrowed heavily to build data centers. While the broader cloud sector shrugs, the most leveraged names are getting hammered.
| Stock | Ticker | Today’s Move | 1-Month Move | Key Detail |
|---|---|---|---|---|
| Nebius Group | NBIS | -10% ($152.58) | -43% | Biggest decliner in group |
| CoreWeave | CRWV | -9% ($61.53) | -36% | CDS implies 50% default odds |
| Oracle | ORCL | -2% ($117.29) | — | Largest non-financial borrower in high-grade index |
| NVIDIA | NVDA | -3% ($191.59) | — | CDS also at new highs |
| SKYY Cloud ETF | SKYY | +0.59% ($138.71) | — | Broad cloud software holding up |
Simple Explanation: What is a CDS?
Think of a Credit Default Swap (CDS) like car insurance for bonds. If you lend money to a company (buy their bonds), you can buy a CDS to protect yourself if they can’t pay you back.
- Higher CDS cost = investors think the company is riskier
- Basis points (bps): 100 bps = 1%. So 855 bps = 8.55% per year to insure against default.
- Implied default probability: Models translate CDS prices into "odds the company fails within 5 years."
The trigger isn’t bad earnings — it’s the credit market. Investors are asking: "Can these companies afford their massive debts with interest rates this high?"
CoreWeave is the poster child for this selloff. Here’s why the market is worried:
| Metric | Status | Why It Matters |
|---|---|---|
| CDS Cost | ~855 basis points (8.55%/year) | Extremely high — like insuring a house in a hurricane zone |
| Implied 5-Year Default Probability | ~50% | Coin-flip odds the company can’t pay its debts |
| Credit Rating | Junk (below investment grade) | No safety net for bondholders |
| Free Cash Flow | Negative since 2022 | Burning cash, not generating it |
| Q1 2026 Interest Expense | $536 million (doubled) | Debt payments eating revenue |
| Q1 2026 Free Cash Flow | -$4.71 billion | Massive cash burn |
Important Callout
CoreWeave’s financials show a company spending aggressively to win the AI infrastructure race — but the bill is coming due. With negative cash flow and doubling interest costs, the market is questioning whether the business model works at current interest rates.
Here’s the most important chart of the day: SKYY Cloud ETF is GREEN (+0.59%).
What This Tells Us
SKYY holds 3% CoreWeave and 3.8% Oracle — both dragging it down — yet the ETF still rose.
→ Broad cloud software (SaaS, applications, platforms) is healthy.
→ The pain is CONFINED to leveraged AI infrastructure builders, not the whole cloud sector.
→ This is a targeted re-rating, not a cloud-wide crash.
Recent filings show big insider sales — but context matters:
| Insider | Company | Amount Sold | Plan Type | Context |
|---|---|---|---|---|
| Brian Venturo (Co-founder) | CoreWeave | ~$734 million | 10b5-1 | Pre-arranged, routine wealth management |
| Michael Intrator (CEO) | CoreWeave | ~$447 million | 10b5-1 | Pre-arranged, routine wealth management |
| NVIDIA Director | NVIDIA | ~$407 million | 10b5-1 | Pre-arranged, routine wealth management |
ELI5: What is a 10b5-1 Plan?
A pre-written schedule to sell shares automatically at set times/prices.
- Set up months in advance when insiders don’t have secret info
- Prevents accusations of insider trading
- Standard practice for executives diversifying wealth
- Not a panic signal — executives still hold large remaining stakes
CDS Spreads at Close
→ If Oracle, CoreWeave, Nebius CDS stay wide (high), refinancing gets harder → equity multiples compress further
SKYY ETF Holds Green
→ Confirms pain stays confined to leveraged AI buildout names
NVIDIA Stock Stabilizes
→ Bellwether for broader AI sentiment
Fresh Financing Announcements
→ Any new debt/equity from CoreWeave or Nebius → sets tone for Thursday’s open
| Topic | Bottom Line |
|---|---|
| Today’s Move | Leveraged AI infrastructure stocks (NBIS, CRWV) crashed 9–10% |
| Root Cause | CDS costs surged → market prices in ~50% default odds for CoreWeave |
| Broader Market | Not a cloud crash — SKYY ETF up, software names fine |
| Key Risk | High rates + negative cash flow = refinancing wall for heavy borrowers |
| Insider Sales | Pre-planned (10b5-1), not panic — executives keep big stakes |
| Next Catalyst | CDS levels at close, any financing news, Treasury yields |
The AI buildout isn’t stopping — but the "free money" era is over. Companies that borrowed heavily to build GPU clusters now face a math problem: can they earn enough rent on those GPUs to cover rising interest bills? The market just answered "maybe not" for the most leveraged players.
Not necessarily. NVIDIA’s CDS rising reflects systemic worry about AI capex financing, not NVIDIA-specific trouble. NVDA has massive cash flow, low debt, and pricing power. The CDS move is likely contagion from weaker names. Watch if NVDA’s fundamentals change.
It’s a model-derived estimate based on CDS pricing — not a prediction. It means: "If you insure CoreWeave’s debt for 5 years at today’s price, the market thinks there’s a coin-flip chance they miss a payment." It could be wrong. But it does make new borrowing very expensive.
SKYY holds ~70+ stocks. CoreWeave (3%) and Oracle (3.8%) are small weights. The other 93% (Microsoft, Salesforce, ServiceNow, etc.) are rising, outweighing the drag. This proves the selloff is selective, not sector-wide.
No. 10b5-1 plans are set up months ahead by lawyers/compliance teams. The CoreWeave insiders still own huge stakes — they’re diversifying, not exiting. The NVIDIA director sale is similarly routine.
Three things:
Article based on market data from July 29, 2026. For informational purposes only — not investment advice. Consult a financial advisor before making investment decisions.