Core Scientific's $600 Million Credit Deal Is a $300 Million Cash Unlock

On August 27, Core Scientific (CORZ) told investors it had secured $600 million in new senior secured credit facilities. The headline number was precise, committed, and bank-backed by JPMorgan, Morgan Stanley, and Goldman Sachs. The market's reply was a shrug. Shares came under pressure in the wake of the announcement, and the stock drifted lower again on Tuesday, near $18, against a market value around $5.8 billion.

The cool reception is the clue worth following. "Secured $600 million" sounds like a company being handed a fresh war chest. Read the documents and the number is not what it appears. The facilities are a $100 million revolving credit line and a $500 million letter of credit facility. A letter of credit is not new capital; it is a bank promise that replaces cash Core Scientific had already posted as collateral. The real, trackable effect sits one line down in the release: chief financial officer Jim Nygaard said the deal is expected to release about $300 million of restricted cash.

That gap between the headline figure and the actual unlock is the story works until line. The $600 million headline is best read as roughly a $300 million working-capital release, wrapped in the language of a credit win. Investors who took the announcement at face value as "more money to grow" were reading a different deal than the one on the page.

The number hiding inside the headline

Start with what a letter of credit facility does. When a data center operator signs a power agreement with a utility, the utility often wants assurance it will be paid for capacity reserved years out. In the past, Core Scientific satisfied that requirement by posting cash. That cash sat restricted on the balance sheet — real money, locked up, earning nothing, not available for the build.

The new facility swaps that structure. Instead of cash sitting as collateral, a bank issues the letter of credit, and the company pays fees on it (1.75% annually on outstanding amounts, plus a fronting fee). In return, roughly $300 million of previously restricted cash comes back into the company's control. This is legitimate treasury work, not a red flag. Cash collateral is an inefficient use of balance sheet; substituting a committed bank facility for it is what a well-managed borrower does. The mechanism checks out.

The reason the market did not cheer is what the mechanism does not fix. Freeing $300 million matters at the margin precisely because this is a company whose cash is doing double duty. As of June 30, Core Scientific reported total liquidity of about $1.82 billion in cash and digital assets at the same time it carried $7.68 billion in total liabilities and a $2.42 billion stockholders' deficit. That is a negative-equity balance sheet financing an enormous construction program.

A build-first capital structure

The credit facility has to be read against how Core Scientific is funding its pivot from bitcoin miner to AI data center landlord. The company emerged from an earlier chapter with a plan to repurpose its power-rich sites into high-density colocation (HDC) for AI workloads, and it is spending ahead of contracts. In the first half of 2026 it raised roughly $4.28 billion in debt proceeds, paid down about $1 billion of prior obligations, and poured $954 million into property and equipment plus another $232 million into land and development rights. Operating cash flow over the same six months was only about $231 million.

The gap between what the build costs and what the operations generate is bridged with debt. So the reported profit-and-loss is dominated not by the selling of anything but by non-cash mark-to-market swings: the second quarter's $1.16 billion net loss was driven almost entirely by a $1.05 billion non-cash change in the fair value of warrants. Adjusted EBITDA was a real $41.1 million, but the accounting swings — and the warrant overhang they point to — keep reminding shareholders that a meaningful slice of the equity claim is tied to future dilution.

The revenue story is real, and it is the reason the stock trades at roughly 13 times trailing sales instead of being treated as a distressed miner. Colocation revenue jumped to $136.7 million in the second quarter, up from $10.6 million a year earlier. The company is billing 437 megawatts of capacity, which it says annualizes to about $635 million of colocation revenue, and it holds leases on about 1.1 gigawatts. It recently anchored an AMD partnership with fifteen-year agreements for roughly 530 megawatts across five sites, representing more than $14 billion of potential contracted revenue.

That is the bull case, and it is real. The tension is timing and cash. Data center revenue is recognized when capacity comes online and bills, but the cash to build that capacity is spent years earlier. The $300 million the credit facility releases is a cushion, not a fix — it buys room in a model that must keep raising and spending to convert its contracted pipeline into billed megawatts.

The invoice the facility quietly adds

There is a cost attached to the new liquidity, and it shows up in two places. First, these are senior secured facilities secured by a first-priority lien on substantially all of the company's assets, guaranteed by its domestic subsidiaries, and carrying maintenance covenants that Core Scientific must now live under: a minimum liquidity of $150 million and a minimum collateral value of $3 billion. The company is effectively telling its lenders, and itself, that with $1.82 billion of liquidity and a $7.68 billion debt load, the margin of safety that matters has to be watched.

Second, the deal is a confession of the funding loop. Growth needs power, power needs collateral, collateral ties up cash in a company that is already negative-equity and free-cash-flow negative while it builds. Each turn of that crank adds a secured claim at the top of the capital stack, and the securities that have absorbed the risk — the warrant and contingent-value paper showing up as fair-value swings on the income statement — are the shape of the dilution that funds the rest.

The core confusion the reader should resolve is simple. "Core Scientific secured $600 million" is not a vote of new confidence in the AI story; it is a ≈$300 million efficiency that lets a stretched balance sheet keep building. The banking muscle behind it — JPMorgan, Morgan Stanley, Goldman — is real and worth something. So is the contracted revenue, if it comes online. What the credit deal does not do is change the arithmetic: this company must keep tapping debt and diluting equity until the 437 billed megawatts catch up with the billions already spent.

The next document that moves this case is not the credit agreement. It is the quarterly billing progress — whether the 1.1 gigawatts of contracted capacity turns the $1.8 billion of liquidity into the kind of operating cash that stops the need for the next facility. Until then, treat the $600 million announcement as what it is: headline force around a $300 million unlock, at a company that needed it more than the press release let on.