Core Scientific is dirt cheap at under 10x contracted annual net operating income, given that most of that income now comes from investment-grade AMD. x.com 

Core Scientific has more than 4 gigawatts to energize and $500 million of free cash flow coming from CoreWeave. Selling because it announced no new deal with earnings is foolish, since companies in this group rarely announce deals alongside earnings. x.com 

Core Scientific has stopped being a bitcoin miner and become a contracted AI-infrastructure landlord with years of revenue visibility. Its Q2 loss is mostly a non-cash warrant mark, not a broken business. AMD signed as anchor tenant for over 500 MW from 2027, scalable to 2.5 GW, earning warrants only as capacity comes online. The test is on-schedule delivery into 2027; if AMD and CoreWeave both scale as contracted, this is one of the more de-risked demand stories in the AI power trade. x.com  … Show more

Core Scientific looks attractive as a catch-up trade. x.com 

Galaxy Digital and Core Scientific should outperform Applied Digital over the next five years; buying Applied Digital now because it has surged looks foolish. A year ago it was a near-bankrupt bitcoin miner; it has since signed 900MW of colocation deals, a levered, lower-quality business rallying on an industry trend, like Intel versus Nvidia. Galaxy's profitable digital-assets arm and Core Scientific's earlier deals meant no bankruptcy risk was priced in, so they lagged. This could be wrong. … Show more

Core Scientific's selloff over doubts about CoreWeave misses that, if CoreWeave cannot meet its obligations, Core Scientific can re-lease that capacity at 40-60% higher rates. Most of its capacity is tied to CoreWeave, which is why the stock is being hit as CoreWeave's viability is questioned. x.com 

Core Scientific's big lease with AMD, 500MW to start with a path to 2.5GW, looks good on the math. Per the call, it brings $14B for 530MW over 15 years, 380MW of it triple-net. AMD receives market-priced warrants to buy Core Scientific equity. x.com  x.com 

Core Scientific is the lowest-risk of these AI infrastructure picks, ranked third, because CoreWeave funds the buildout and pays rent under take-or-pay contracts whatever happens in the compute market. It has 590MW contracted, with capex borne by the customer, so it needs little debt. Nvidia's Rubin does not hurt it, since CoreWeave would swap GPUs but keep using its sites and power. The only risk is total dependence on CoreWeave: any trouble there hits Core Scientific directly. x.com 

CoreWeave looks set to test its all-time highs as it bounces off the 0.618 Fibonacci retracement, with large positive gamma and vanna exposure concentrated at the $110 strike for 18 September. x.com 

CoreWeave is a name to avoid because the neocloud model is financial engineering and insiders are hedging. Its CFO sold 30-40% of his shares. @AtlasShrug1 adds that CoreWeave's capital structure is a mess and puts it in the high-risk-of-permanent-capital-impairment bucket despite its large market cap. x.com