By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel is bearish on Core Scientific's (CORZ) ambitious data-center expansion, with key concerns being grid interconnection delays, labor scarcity, and financing risks that could derail the project's timeline and economics.
Risk: Grid interconnection delays and the potential forfeiture of revenue recognition due to penalty clauses in the contracts.
Opportunity: None explicitly stated in the discussion.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Core Scientific (NASDAQ:CORZ) said it has entered a commercial partnership with AMD covering up to 2.5 gigawatts of data center capacity, beginning with approximately 530 megawatts across five campuses under 15-year agreements.
<pre><code> Chief Executive Officer Adam Sullivan said the initial agreements represent more than $14 billion of base contracted revenue and include 2.5% annual escalators. The company now has about 1.1 gigawatts of total contracted billable capacity, representing more than $24 billion of base contracted revenue, according to management. → Volatility Is Back and These 3 Market Tollbooths Are Best Positioned to Profit The AMD commitment includes roughly 380 megawatts delivered directly to AMD through triple-net leases at Core Scientific's Pecos, Hunt and Muskogee sites. The remaining approximately 150 megawatts at the Auburn and Dalton campuses will support an unnamed Neocloud customer through modified gross leases, with AMD providing full credit support for the 15-year terms. Sullivan said the AMD credit support arrangements do not include equity step-in rights, which he said protects Core Scientific's equity investments in the projects. The company also issued AMD a warrant with a strike price reflecting current market levels, subject to certain commercial conditions. ## Expansion rights and development plans → 2 Stocks Built to Thrive If Inflation Refuses to Fade AMD has exclusive reservation rights, at specified times and under specified conditions, to lease as much as an additional 2 gigawatts of capacity. Core Scientific said it believes it can make that capacity available through additional grid-connected power, capacity progressing through load studies and behind-the-meter power solutions at Pecos, Hunt and Muskogee. Management did not provide specific timing for the additional capacity. Sullivan told analysts that the company would work closely with AMD on future requirements and would not discuss capacity covered by the reservation agreement with other prospective customers. → This Tiny AI Supplier Could Be More Important Than the Chipmakers The company said Pecos and Muskogee each have the potential to support up to 1 gigawatt of leasable capacity through a combination of additional grid power and behind-the-meter solutions. During the question-and-answer session, management said behind-the-meter development timelines could be driven by lateral pipeline construction and equipment availability, depending on the site. The company is considering technologies including turbines, fuel cells, linear generators and other alternatives. Separately, Core Scientific said it has identified a pipeline of more than 2 gigawatts of potential incremental power at new sites, with initial capacity potentially available from late 2028 through 2030. The company said it intends to selectively acquire powered land and develop new locations while continuing to execute contracted projects. ## Construction milestones and delivery schedule Chief Operating Officer Matt Brown said Core Scientific is currently billing for 437 megawatts of capacity, having substantially completed four of five CoreWeave campuses ahead of its prior timeline. The company had expected to substantially complete those sites before the end of summer. Dalton Phase Two, the fifth and final CoreWeave campus, is expected to begin delivering its remaining 150 megawatts at the end of 2026 and be completed in early 2027, Brown said. For AMD, Pecos is expected to be the first location to begin delivery, with initial megawatts scheduled for the first half of 2027. Management said about half of the 530-megawatt commitment is expected to be delivered in 2027, with the balance delivered in 2028. Brown said vertical construction is underway at Pecos, major infrastructure equipment is arriving, and the building shell for the first 185 megawatts is nearing completion. Hunt, Auburn, Muskogee and Dalton Phase Three are advancing through design, procurement, site preparation and construction. Management said the company had begun developing several AMD-related locations before customer contracts were signed. It had secured long-lead equipment, released capital for substation construction at multiple sites and engaged general contractors and labor resources. Sullivan said the company has equipment secured and contractors on site across the five campuses. Core Scientific expects all-in construction costs of approximately $11 million to $12 million per megawatt for high-density AI infrastructure. Brown said that estimate includes construction labor, owner-furnished equipment such as transformers, switchgear, generators and cooling systems, as well as engineering, permitting, utility interconnection, insurance, testing and other soft costs. He said higher costs relative to the earlier CoreWeave program reflect increases in equipment and labor costs, supply-chain constraints, tariffs and differences in site design. Management said labor availability remains a significant construction cost driver in competitive U.S. markets. ## Second-quarter financial and funding update Chief Financial Officer Jim Nygaard said second-quarter GAAP colocation revenue rose sequentially to $137 million, aided by the increase in billable capacity. He said the company expects another meaningful increase in colocation revenue during the third quarter. Under GAAP accounting, revenue from the CoreWeave contracts is recognized on a straight-line basis over the 12-year lease terms, effectively recognizing contractual escalators earlier in reported revenue, Nygaard said. Core Scientific continued to wind down its Bitcoin mining operations during the quarter. The company ended June with nearly 30% fewer miners online than at the end of the first quarter and was self-mining at only two sites. Nygaard said mining activity is expected to continue declining through the remainder of 2026 as the business is operated primarily to offset contractual power costs. Second-quarter cash selling, general and administrative expense was about $36 million, up $4 million sequentially, primarily due to one-time professional fees related to recent debt financing. Nygaard said the company views the low-$30 million range as a reasonable quarterly SG&A baseline, though expenses may vary as it supports growth. Core Scientific ended the second quarter with approximately $1.8 billion in liquidity. The company estimates the initial AMD build-out will require about $6 billion of capital and expects to finance it through project-level bonds. It also said it is prepared to invest up to approximately $1 billion to advance roughly 500 megawatts of initial future capacity ahead of customer contracts. Nygaard said the company expects to use project-bond structures similar to those used for its CoreWeave projects. He added that the direct AMD leases and the Neocloud leases backed by AMD credit support would have distinct contractual structures but would be financed through similar project-level vehicles. ## About Core Scientific (NASDAQ:CORZ) Core Scientific, Inc (NASDAQ: CORZ) is a leading provider of large-scale blockchain infrastructure and digital asset mining services. The company develops, owns and operates high-performance data centers optimized for the mining of Bitcoin and other proof-of-work cryptocurrencies. In addition to its core mining operations, Core Scientific offers colocation, hosting and managed services designed to support institutional clients and enterprise users in deploying and scaling blockchain nodes and computing hardware. Core Scientific's service portfolio includes hardware procurement, deployment and maintenance, real-time monitoring, power management and network connectivity. *This instant news alert was generated by narrative science technology and financial data from MarketBeat in order to provide readers with the fastest reporting and unbiased coverage. Please send any questions or comments about this story to [email protected].* The article "Core Scientific Q2 Earnings Call Highlights" was originally published by MarketBeat. **View MarketBeat's top stocks for July 2026****.** </code></pre>Four leading AI models discuss this article
"The multi-year, inflation-linked revenue visibility from the AMD partnership outweighs near-term mining wind-down and elevated build costs, supporting re-rating of CORZ as an AI infrastructure play."
CORZ's AMD deal adds $14B+ in 15-year contracted revenue with 2.5% escalators across 530 MW initially, lifting total contracted billable capacity to 1.1 GW and $24B. Delivery starts 1H27, construction costs are elevated at $11-12M/MW, and the company is pivoting hard from Bitcoin mining (down 30% QoQ) toward AI/HPC colocation. Liquidity of $1.8B and project-level bond financing provide runway, but timelines slip into 2028-2030 for incremental GW-scale capacity. The warrant issuance and AMD credit backing reduce equity dilution risk yet still embed meaningful counterparty concentration.
The article glosses over execution risk on behind-the-meter power solutions (turbines, fuel cells) whose timelines depend on pipelines and equipment that have repeatedly delayed similar projects; $6B capex for the AMD build-out assumes project bonds close on favorable terms in a higher-for-longer rate environment, and any slippage in 2027-2028 deliveries could trigger revenue recognition delays or covenant pressure.
"Core Scientific has successfully de-risked its business model by pivoting to long-term, credit-backed infrastructure leases, effectively becoming a proxy for AI-driven power demand."
Core Scientific (CORZ) is aggressively pivoting from a volatile Bitcoin miner to a high-moat infrastructure utility for AI hyperscalers. The $24 billion in contracted revenue provides massive visibility, shifting the valuation model from commodity-linked hash rate to long-term, inflation-protected cash flows. However, the $11-12 million per megawatt construction cost is steep, and the reliance on project-level bonds creates significant leverage risk. While the AMD partnership validates their technical capability, CORZ is now essentially a massive, capital-intensive construction firm. Investors must watch if they can maintain margins against rising labor costs and potential grid interconnection delays that could derail the 2027-2028 delivery schedule.
The company is trading its operational agility for massive, long-term debt obligations, leaving them highly vulnerable if AI demand cools or if they face cost overruns on their $6 billion capital expenditure plan.
"AMD's 15-year, $14B commitment with 2.5% escalators de-risks CORZ's capex model, but execution on $11–12M/MW across five concurrent builds in tight labor markets is the real hurdle."
CORZ has locked in $24B of contracted revenue across 1.1 GW, with AMD's 15-year, $14B commitment providing visibility. The 2.5% annual escalators and project-level financing (avoiding balance-sheet dilution) are structurally sound. However, the real test is execution: $11–12M/MW capex on five simultaneous builds, labor scarcity in competitive U.S. markets, and 2027–2028 delivery windows create binary risk. AMD's 2 GW expansion option is optionality, not revenue. Mining wind-down is prudent but masks that CORZ still depends on colocation ramp to justify $6B capex.
The $11–12M/MW cost estimate assumes no further supply-chain shocks, tariff escalation, or labor inflation—all explicitly cited as headwinds. If actual costs drift to $13–14M/MW, project-level financing becomes unviable and CORZ absorbs equity losses.
"If Core Scientific can execute the AMD-backed build-out on time and finance it via project bonds, the company could achieve highly visible, long-duration revenue scaling that may justify a material re-rating, even as execution and financing risks loom."
Core Scientific's Q2 highlights a pivot from crypto mining toward a capital-intensive, AMD-backed data-center platform. The 530 MW initial tranche on 15-year leases, plus options for up to 2 GW, signals a durable, revenue-rich roadmap far beyond traditional mining. If execution stays on plan, the combination of base contracted revenue (>$14B on 530 MW; >$24B on 1.1 GW) and long-term escalators could support a meaningful re-rating via project-bond financing and AMD credit support. Yet the plan hinges on massive capex (~$6B initial), grid interconnections, labor, supply chains, and steady demand—risks that could delay or erode returns.
The strongest counter is execution risk: project delays, interconnection bottlenecks, and cost overruns could push cash flows further out, while higher financing costs in a rate environment could erode returns despite the long-dated contracts.
"Behind-the-meter power permitting timelines represent the largest unpriced execution risk beyond labor and capex inflation."
Claude's labor-scarcity point is real but incomplete: the bigger unmentioned risk is behind-the-meter generation permitting. Turbines and fuel cells face multi-year environmental and interconnection queues that have already pushed similar 100+ MW projects past 2028. If those slip, the $6B capex timeline collapses regardless of AMD credit or bond terms.
"The market underestimates the lethal impact of performance-based penalty clauses in these hyperscaler contracts if construction timelines slip."
Grok and Claude are fixated on construction and permitting, but both ignore the 'take-or-pay' mechanics of these contracts. If CORZ fails to deliver power on time, the penalty clauses could be lethal. These aren't just construction contracts; they are performance-based utility agreements. If interconnection delays occur, the revenue recognition isn't just delayed—it's potentially forfeited. We are treating this as a construction project when it is actually a massive, binary operational liability.
"Take-or-pay penalties matter less than cost-overrun exposure when capex estimates assume no tariff or labor escalation beyond 2024 baselines."
Gemini's take-or-pay penalty framework is sharper than I credited. But the contracts likely have force majeure clauses for grid interconnection delays—standard in utility PPAs. The real bite is whether CORZ absorbs cost overruns mid-build. If a turbine permit slips 18 months, AMD's contract probably survives; if capex balloons 20%, CORZ eats it. That's the asymmetry nobody's quantified.
"Financing risk and rate shock threaten the whole project – even with take-or-pay and AMD backing, higher funding costs and cost overruns could push cash flows out of reach and trigger covenants long before 2027–2028 deliveries."
Gemini's focus on penalties covers downside, but the bigger flaw is financing risk baked into the plan. If capex runs $11-12M/MW and interconnections slip, project bonds must refinance at worse terms in a higher-for-longer rate environment. That raises cash-costs and could push revenue recognition out of sync with debt service, triggering covenants even with AMD backstops. The binary risk isn't just delays; it's whether the IRR survives elevated financing costs and potential overruns.
The panel is bearish on Core Scientific's (CORZ) ambitious data-center expansion, with key concerns being grid interconnection delays, labor scarcity, and financing risks that could derail the project's timeline and economics.
None explicitly stated in the discussion.
Grid interconnection delays and the potential forfeiture of revenue recognition due to penalty clauses in the contracts.