The Texas-based company (NASDAQ: CORZ) secured the 364-day loan at SOFR plus 250 basis points, a notably tight spread for a firm with roots in the volatile crypto mining sector. The deal includes an accordion feature (a provision allowing the borrower to draw additional commitments under agreed conditions) that could raise total commitments to $1 billion.

Proceeds are earmarked for equipment purchases, real property acquisition, pre-development costs, and energy procurement agreements tied to high-density AI data centers. Core Scientific's total liquidity stood at $530 million at the end of 2025, prior to the Morgan Stanley closing.

CORZ shares slipped roughly 1% in pre-market trading on the announcement day, suggesting the market had partially priced in the news.

The deal came three days after the company reported its Q4 2025 earnings on March 2. The proximity to the earnings release may reflect a deliberate signal of financial confidence following those results, though the company has not commented on the timing.

---

Mining Is Being Wound Down

Core Scientific CEO Adam Sullivan has been direct about the company's direction. Its Bitcoin mining segment is, in his words, "essentially in runoff," with mining sites being converted to colocation facilities that host AI and high-performance computing workloads. The company sold 1,900 BTC for approximately $175 million in January 2026 and now holds fewer than 1,000 BTC on its balance sheet. It plans to monetize "substantially all" remaining Bitcoin reserves this year to fund the infrastructure buildout.

The revenue shift is already visible in the numbers. Q4 2025 mining revenue came in at $41 million, while colocation services generated $31 million. That gap has been narrowing.

By end-2028, the company says every megawatt in its portfolio will be dedicated to AI colocation, with a target of 1.5 gigawatts of leasable power capacity by that date.

Sullivan framed the Morgan Stanley financing in operational terms. "This strengthens our liquidity and enhances our financial flexibility as we execute our development and go-to-market strategy," he said in the official announcement. He told Decrypt: "With this additional financing capacity, we can operate decisively by deploying capital to expedite project ready-for-service timelines."

---

One Customer, One Very Large Contract

The company's AI revenue is heavily concentrated in a single client. CoreWeave, a GPU cloud provider, accounts for approximately 76% of Core Scientific's projected 2026 revenue under a 12-year, $10 billion-plus take-or-pay contract. That type of contract obligates both parties to fulfill their commitments and cannot be unilaterally terminated. By the end of 2025, 213 megawatts were energized under the arrangement, with roughly 100 MW actively billing. The full 590 MW build-out is targeted for completion by early 2027.

That level of client concentration represents a structural risk worth naming directly. A single counterparty accounting for roughly three-quarters of projected revenue is a meaningful variable in any institutional lender's underwriting calculus. Morgan Stanley's willingness to extend credit at relatively tight spreads implies confidence in CoreWeave as a counterparty, but the concentration is a factor investors and creditors should monitor.

Compass Point analysts maintained a Buy rating on CORZ with a $28 price target following the Morgan Stanley announcement, citing confidence in management's execution and its ability to attract investment-grade customers.

---

Wall Street Is Now Writing Checks to Crypto Infrastructure

Core Scientific is not the only former miner making this transition. Sector-wide capital expenditure on data center infrastructure increased 400% between March 2025 and February 2026, according to Insights4VC, a crypto infrastructure research newsletter.

Hut 8 secured 85% loan-to-cost financing from JPMorgan and Goldman Sachs for a $7 billion Google-backed AI project, according to Insights4VC, underscoring how mainstream lenders are now underwriting compute infrastructure that originated in the crypto space.

For firms that have locked in AI contracts, mining is projected to fall from roughly 85% of total revenue to under 20% by the end of 2026, Insights4VC reports.

---

What This Means Outside the United States

For developers and infrastructure operators in South Asia and Sub-Saharan Africa, the Morgan Stanley deal is a useful reference point for what their markets currently lack.

India controls only 3% of global data center capacity despite hosting nearly 20% of the world's data. The country needs an estimated $25 billion to $45 billion in new investment to add around 8 gigawatts of capacity by 2030. Sixty to eighty percent of existing Indian data centers face high water stress, and land acquisition and regulatory coordination remain fragmented. A separate measure sharpens the investment gap further: India receives just 1.12% of global AI investment despite representing 6.6% of global GDP. Hyperscalers are moving to close that gap on the infrastructure side, with Microsoft committing $17.5 billion, Google $15 billion, and AWS $12.7 billion to Indian data center buildouts, but structured debt financing of the kind Morgan Stanley provided to Core Scientific has not followed at comparable scale.

The capital architecture behind the Core Scientific deal, a large institutional lender extending purpose-built credit to a dedicated compute operator at tight spreads, does not yet exist at scale in South Asian markets.

Africa's situation is starker. The continent operates roughly 360 megawatts of data center capacity, which is about 0.6% of worldwide installed capacity. Its five largest markets are projected to grow to 1.5 to 2.2 gigawatts by 2030, supported by deals including the IFC's $100 million commitment to Raxio spanning six African countries and a $700 million GPU facility programme by Cassava Technologies. A Naver-NVIDIA-Nexus consortium is separately developing a 500 MW AI campus in Morocco, the continent's most prominent current AI infrastructure project. South Africa's trajectory illustrates the pace of regional growth: its AI data center market is forecast to expand from $70 million in 2025 to $572 million by 2031, a compound annual growth rate of 42%.

But equivalent TradFi willingness to write billion-dollar loan facilities to African compute operators has not yet materialised at scale.

CoreWeave, the primary beneficiary of Core Scientific's expansion, does serve AI customers globally, including in emerging markets. As its underlying U.S. infrastructure grows, GPU availability and pricing on CoreWeave's cloud platform could improve for developers in Lagos, Nairobi, Mumbai, or Dhaka. In the view of this publication, that represents an indirect downstream benefit rather than a structural solution to local infrastructure gaps.

The more direct takeaway is structural: as compute capacity concentrates in U.S. facilities financed by Wall Street debt, cross-border latency and cost pressures for AI workloads will remain a constraint for builders outside North America. That conclusion is consistent with the broader body of research on compute access as a gating constraint for startups in emerging regions.