Neoclouds such as CoreWeave and Nebius are backstopped by companies with trillion dollar market caps and access to the cheapest financing in the world NVIDIA has a 10% equity stake in CoreWeave and has committed to buy any unused capacity through 2032, up to $6.3B NScale, Lambda, Nebius and Iren have equity ownership and preferential vendor agreements with NVIDIA Microsoft has signed a contract and prepaid 20% of the value at Iren. NVIDIA has backed leases at both NScale and Lambda. Google has backed both Hut8 and TeraWulf through guarantees on the lease. Amazon has guaranteed leases for Cipher. Yet, people still believe that these companies are at risk of default simply from looking at credit default spread charts

I bought more $CRWV this week. The debt deserves scrutiny but I think the market has gotten too pessimistic about CoreWeave’s ability to finance its growth and turn customer commitments into cash. CoreWeave's trailing twelve-month revenue through Q2 went from about $3.5 billion to $7.6 billion which is up roughly 115%. EBITDA on the same basis went from about $1.8 billion to $3.8 billion. That is not the company’s adjusted EBITDA and it is definitely not free cash flow. What it shows is how fast the operating business is scaling. The real question is whether that growth earns a good return after paying for the infrastructure and the financing behind it. I think it does. CoreWeave ended Q2 with about $104 billion of backlog, and that does not count the more than $25 billion of new commitments it signed early in Q3. Backlog still has to be delivered and collected on, but it gives lenders something real to underwrite. The model management laid out on the Q2 call is simple. Build capacity around a signed contract and let the customer payments cover operating costs, interest and principal. A typical five-year deal is expected to pay off its asset level debt during the original term and throw off extra cash. Anything the cluster earns after that is upside. Management still has to prove it, but that is the model I am buying. The older hardware is holding up too. CoreWeave signed an A100 contract that runs into 2029, and the A100 came out in 2020. I am not saying every GPU stays profitable for eight years. I just think the idea that older chips become worthless overnight is wrong. When I look at the bonds the 9.625% notes due 2032 still pay 9.625%. A lower price and a higher yield do not change the coupon on debt that is already outstanding. The selloff does matter because it signals more concern about repayment and makes new unsecured debt more expensive. If funding costs climb while customer pricing is locked in, returns get squeezed, and that is a fair risk to raise. What I would not do is treat a stressed bond yield as the cost of every future dollar they borrow. In March CoreWeave closed an $8.5 billion facility with a floating tranche at SOFR plus 2.25%. In August it added $2.6 billion at SOFR plus 5.50%. In September it raised $4.2 billion of converts at a 2.875% coupon. None of that is a quote on the next deal. It shows that customer quality, collateral and structure drive very different borrowing costs. It also does not make the parent bulletproof, since roughly 90% of June debt was recourse. The cash that CoreWeave needs to produce is immense due to their debt structure. In the first half CoreWeave generated about $3.7 billion of operating cash flow while spending $14.1 billion on CapEx while repaying $5.2 billion of principal. At this rate CoreWeave will still need outside money. Management can close the gap by finishing funded projects and collecting on them, lining up financing before taking on new commitments, pushing for customer prepayments and using equity or converts when the terms make sense. Prepayments are already part of it as roughly 70% of the deals signed in Q2 included some upfront payment. Some of these choices mean dilution or slower growth, and I would rather see management protect liquidity and project returns than chase the biggest revenue number. What would change my mind is persistent deployment delays, weak collections, financing costs eating too much of the return or dilution that outruns the value being built. I bought more because I think the return justifies those risks. Now I want to see projects completed, cash collected and financing terms that leave enough on the table for shareholders.