Goldman Upgraded CoreWeave. Nebius Got the Bounce. Why the Neocloud Split Is About Balance Sheets, Not Demand

Written Tuesday 20 Oct, SGT, using Monday's US closes.

What happened on Monday

Goldman Sachs analyst Gabriela Borges upgraded CoreWeave to Buy from Neutral. Usually that's enough to move a stock that's down 25% in a week. It wasn't.

- CoreWeave: up about 2.2% intraday to $62.74, then reversed to close down 0.44% at $61.12

- Nebius: closed up 2.88% at $170.24, after rising as much as 4.64% intraday on a reiterated Truist Buy rating with a $355 target

That's a gap of about 3.3 percentage points in one session between two companies that, a week earlier, traded as the same stock. Last week CoreWeave fell 25.21% and Nebius fell 25.15%, almost exactly the same.

So the split started on Monday. The question is why, and whether it keeps widening.

Read Goldman's note carefully

The headline says "Buy", but the details are more mixed:

- The price target was cut, from $139 to $106. That's still about 73% above Friday's $61.39 close.

- The valuation multiple was lowered, from 22x to 18x estimated EBIT for the fifth through eighth quarters out.

- The core argument: investors overestimate CoreWeave's future cost of capital and underestimate how much its returns can improve. Borges says the company's value depends on the spread between its unlevered return and its cost of capital.

- Pricing power: CoreWeave raised prices 25% in July and another 10% in October.

- Goldman still flags financing at an acceptable rate as a key risk.

In short, Goldman thinks the stock is too cheap but is less confident about how much it's worth. When the upgrade comes with a 24% target cut, the market reads both halves.

The real reason for the split: debt

Both companies sell the same thing: GPU capacity to AI labs and hyperscalers. Their balance sheets are completely different.

CoreWeave:

- Total debt: $51.61 billion

- Cash: $5.54 billion

- Net debt: about $46.1 billion

- Market cap: about $33.86 billion

- Debt is about 1.52 times the stock market value

- Q3 interest expense guidance: $860 to $940 million

- Five-year CDS closed Friday at 1,117 basis points

Nebius:

- Total debt: $10.20 billion

- Cash: $8.04 billion

- Net debt: about $2.2 billion

- Market cap: about $46.66 billion

- Debt is about 0.22 times the stock market value

That's the whole story. When credit markets get nervous, the company that owes 1.5x its equity value gets punished. The one that owes 0.2x gets the benefit of the doubt. Bond investors are deciding which neocloud can keep growing without needing their permission.

The bear points are real for both

Nebius isn't risk-free either:

- Cash burn: in Q2, operating cash flow was $2.25 billion against $5.66 billion of capex, so operating cash covered only about 40% of spending. That leaves a gap of roughly $3.4 billion it still has to fund.

- Customer concentration: about $32 billion of its $40 billion-plus contracted backlog comes from Microsoft and Meta.

- Capex: $20 to $25 billion planned for 2026.

- Short interest: 16.31% of shares outstanding.

And the issues the topic raises apply to the whole group:

- Depreciation: GPUs lose value as each new generation ships. If the useful life turns out shorter than the accounting assumes, profits are overstated today.

- Repeated fundraising: every new debt or equity raise either costs more interest or dilutes existing shareholders.

- Leverage: when rates rise, as they are now with December hike odds at 74.3%, the most leveraged balance sheet feels it first.

The bull case

- Demand is real. CoreWeave's backlog was $104 billion at the end of Q2, and Nebius has more than $40 billion contracted with investment-grade customers.

- Pricing is rising. CoreWeave's 25% and 10% price increases show compute is still scarce.

- Asset-backed financing tied to contracts means much of the debt is secured by customer commitments, not just hope.

To answer the topic

- Track CoreWeave's financing: yes, because that's what is setting the price right now. Watch the CDS against 1,000bp, the 2031 and 2032 bond yields, and the Q3 interest expense.

- Track Nebius's orders: yes, but the bigger risk there is concentration. Watch whether new customers beyond Microsoft and Meta grow as a share of the backlog.

- Favor steadier cash-flow platforms: in a credit scare, yes. Cash-generating cloud businesses don't need the bond market to keep building.

My view: the split widens as long as credit stays tight. It narrows when the CDS comes down, not when analysts upgrade.

How I apply my sell-put rules to both

1. The 200-day moving average.

- CoreWeave at $61.12 against a 200-day average of $92.44, about 34% below. Hard fail.

- Nebius at $170.24 against a 200-day of $174.47, only 2.4% below. It still fails, but it's close. A reclaim of the 200-day is the first thing I'm watching.

2. The earnings calendar.

- Nebius reports Nov 10, before the open.

- CoreWeave is expected around Nov 9 (not yet confirmed).

- My 30 to 45 day expiries run from about Nov 19 to Dec 4, so every one of them straddles both reports. Both are blocked.

3. If Nebius clears both gates, here is what the strike math would look like, using Friday's implied volatility of about 92% for the Nov 27 expiry (38 days):

- One standard deviation floor: 170.24 minus 170.24 x 0.923 x the square root of 38/365 = about $120

- Moving average anchor: both the 50-day ($222) and 200-day ($174) are above the price, so there is no usable SMA anchor below it

- The delta at $120 is only about 0.09, below my 0.15 floor, so my rule is to solve upward

- A delta of 0.16 to 0.24 puts the strike around $132 to $144

Those numbers are based on Friday's volatility, not a live chain. I'll recompute after earnings.

4. Size it before you get excited.

One $132 put ties up $13,200 of cash. Nebius isn't a core name, so my cap is 2% of the account. That means one contract needs an account of about $660,000. For most people, that answer settles it.

5. CoreWeave stays on the no-touch list.

With the CDS above 1,000bp, the put premium is mostly payment for default risk. My re-entry checklist needs all three: CDS back below 1,000bp, price back above the 200-day, and the Q3 report out with guidance intact.

6. The exits are set on day one.

A 50% take-profit GTC on fill, a stop at 2.5 times the premium for non-core names, and one roll at most.

My plan

No new neocloud puts this week. If I had to pick one name for after earnings, it would be the one with the stronger balance sheet, but only once it reports on Nov 10, reclaims the 200-day and passes the volatility gates (IV Rank of at least 40%, IV Percentile of at least 45%, IV at least 1.2 times historical volatility). Until then, the CDS tells me more than any upgrade.

Your turn: do you think the CoreWeave and Nebius split keeps widening, or does strong GPU demand pull them back together? Which number are you watching most: CoreWeave's CDS, Nebius's 200-day line, or Meta's capex on Wednesday?

This is my personal trading framework, shared for education. It is not investment advice. Options carry significant risk, including assignment and losses larger than the premium collected. Always verify live chain data and confirmed earnings dates before trading.

$CRWV $NBIS $NVDA $MSFT