Core Scientific Inc designs builds and operates large scale purpose built data centers that support high density colocation services and digital asset mining for both its own account and to a lesser extent third party customers The company focuses on power intensive mission critical computing workloads with an emphasis on artificial intelligence and other high performance computing applications As of December 31 2025 it owned or leased ten data centers across seven U S…
Core Scientific Inc designs builds and operates large scale purpose built data centers that support high density colocation services and digital asset mining for both its own account and to a lesser extent third party customers The company focuses on power intensive mission critical computing workloads with an emphasis on artificial intelligence and other high performance computing applications As of December 31 2025 it owned or leased ten data centers across seven U S states representing approximately 1 4 gigawatts of gross utility power capacity or approximately 920 megawatts of total leasable customer power capacity
Core Scientific generates revenue through three primary streams The company earns digital assets for its own account through self mining operations It provides hosting services to third parties for digital asset mining under recurring electricity based consumption contracts Additionally it delivers high density colocation services to third party customers for AI and HPC operations charging fixed payments based on electric capacity and variable payments on a recurring basis Under these arrangements customers pay for leased customer power capacity regardless of utilization under long term take or pay commitments
The company operates through the following segments Colocation Digital Asset Self Mining and Digital Asset Hosted Mining
• The Colocation segment provides space power cooling facilities operations security and other services to third party customers to support workloads for machine learning and artificial intelligence HDC colocation leases may include all or portions of a data center where customers may also lease office space to support their colocation operations with revenue primarily based on power usage as well as square footage.
• The Digital Asset Self Mining segment generates revenue from the deployment and operation of the company s own large fleet of specialized computers or miners within its owned digital infrastructure as part of a pool of users that solve complex cryptographic algorithms to validate transactions conducted on one or more blockchain networks In the Bitcoin network solving a block results in a reward of bitcoin in a process known as mining.
• The Digital Asset Hosted Mining segment generates revenue through the sale of electricity based consumption contracts for hosting services which are recurring in nature The segment provides a full suite of services to digital asset mining customers including deployment monitoring troubleshooting optimization and maintenance of customer equipment and necessary electrical power repair and other infrastructure services required for customers to operate maintain and efficiently mine digital assets.
In the high density colocation market Core Scientific competes with numerous established data center providers including Aligned Data Centers Compass Datacenters Equinix Inc Digital Realty Trust NTT QTS Switch Inc Vantage Data Centers and CyrusOne Inc Many of these competitors are more established have better brand recognition are well capitalized and are organized to take advantage of certain tax benefits lowering their external cost of capital The company believes its experienced data center and engineering leadership team proven ability to rapidly deliver scalable purpose built data centers and cutting edge energy efficient technologies will enable it to compete favorably within the HDC market Additionally Core Scientific competes with other digital asset miners who have signed contracts with AI HPC customers to convert existing digital mining facilities into HDC facilities including Applied Digital Corporation Cipher Mining Inc Galaxy Digital Hut 8 Corp IREN and Terawulf Inc
The company serves third party customers for high density colocation services supporting AI and HPC operations including hyperscale cloud providers AI and other HPC operators and large enterprises with the operational capability to manage hardware and networking at scale For its digital asset hosted mining operations Core Scientific provides hosting services to third party digital asset mining customers who deploy and manage their own hardware while the company designs builds and operates the underlying infrastructure needed to maintain uptime power availability and cooling requirements
Sector:TechnologySector rationaleCore Scientific's primary revenue streams come from operating data centers for AI/HPC colocation and digital asset mining (crypto mining), both of which fall under the Technology sector's 'Data Centers' and 'Crypto Mining' industries. The company designs and builds the specialized infrastructure to support these computing workloads and earns revenue through hosting services and self-mining rewards.Industries:Data CentersTechnologyPrimaryCore Scientific owns and operates physical data center facilities providing high-density colocation services, power, and cooling for AI, HPC, and digital asset mining customers. Its revenue is derived from leasing customer power capacity and providing the physical infrastructure required for these power-intensive workloads.Crypto MiningTechnologySecondaryThe company operates a 'Digital Asset Self Mining' segment where it uses its own fleet of specialized computers to solve cryptographic algorithms and earn rewards in the form of digital assets like Bitcoin.Classified using BQ-MICSCIK: 0001839341
Investment Thesis
▲ Bull case
Core Scientific is positioned to capture exponential growth in AI infrastructure demand through its fully financed, multi-site expansion strategy, with over $2.9 billion in net proceeds from the CoreWeave bond financing providing immediate capital to accelerate development at Pecos, Muskogee, Hunt County, Dalton Phase III, and Auburn sites without waiting for new lease signings. This proactive capital deployment—securing labor, long-lead equipment, and trade partnerships ahead of contracts—creates a decisive competitive advantage by reducing time-to-ready-for-service (RFS) to 12–14 months, directly addressing the market’s most acute pain point: hyperscalers and AI labs cannot wait 18–24 months for new capacity. The company’s ability to show active construction progress with secured supply chains and labor forces transforms sales conversations from speculative discussions into near-term commitment opportunities, as evidenced by immediate re-engagement from three hyperscalers after exclusivity expired at Pecos and Muskogee. This execution-driven model, validated by five operational CoreWeave sites generating over $350 million in annualized GAAP revenue, allows Core Scientific to monetize its development pipeline faster than pure-play developers who rely on lease signings before breaking ground, turning infrastructure readiness into a self-reinforcing demand accelerator.
The strategic shift to milestone-based exclusivity arrangements—replacing open-ended exclusivity with clear, time-bound milestones—eliminates a major historical drag on asset utilization while preserving negotiation leverage, enabling Core Scientific to return high-value sites like Pecos and Muskogee to market immediately when discussions stall, thereby avoiding costly opportunity costs from idle capacity. This approach, confirmed by management’s statement that they “will not keep high-value assets off the market longer than necessary,” directly addresses investor concerns about prolonged exclusivity periods stalling revenue recognition, and is already yielding results with three hyperscalers re-engaging within days of expiration. Combined with the company’s proven ability to scale sites to 1.5 gigawatts gross power at both Pecos and Muskogee through behind-the-meter gas pipelines, grid expansion, and the Polaris acquisition (440 megawatts), this flexible commercial strategy ensures that development progress never stalls due to contractual bottlenecks, allowing continuous capital deployment and value creation across the portfolio. The behind-the-meter infrastructure, while requiring air quality permitting, offers comparable power economics to grid-supplied power while delivering N+30% redundancy—critical for high-availability AI workloads—and reduces reliance on volatile utility interconnection timelines, a structural advantage in an era of grid congestion and permitting delays.
Core Scientific’s transition from Bitcoin mining to high-density colocation is not merely a wind-down but a strategic asset repurposing that unlocks embedded value in existing power infrastructure, with the company retaining only one or two mining sites by year-end while converting the majority of its Texas and Oklahoma campuses to HDC. This shift is de-risked by the fact that colocation revenue from 243 megawatts already covers operating costs and is expanding margins to 80%-85% cash gross profit—up from the original 75%-80% guidance—based on two years of operational experience with actual contractor deployment and headcount data, not theoretical models. The margin expansion is further supported by the lockbox structure in the $3.3 billion bond financing, which prioritizes operating expenses and debt service before allowing surplus cash flow to fund new projects, creating a self-funding growth cycle where CoreWeave-generated cash directly finances expansion at non-CoreWeave sites. Unlike peers reliant on volatile Bitcoin mining revenue or speculative development, Core Scientific now has a stable, contracted revenue base exceeding $350 million annually (straight-line over 12 years) that provides predictable cash flow to support its $2 billion 2026 capex plan, including $700 million for Hunt County and Polaris acquisitions, turning what was once a cost center into a scalable, high-margin growth engine.
Core Scientific is positioned to capture exponential growth in AI infrastructure demand through its fully financed, multi-site expansion strategy, with over $2.9 billion in net proceeds from the CoreWeave bond financing providing immediate capital to accelerate development at Pecos, Muskogee, Hunt County, Dalton Phase III, and Auburn sites without waiting for new lease signings. This proactive capital deployment—securing labor, long-lead equipment, and trade partnerships ahead of contracts—creates a decisive competitive advantage by reducing time-to-ready-for-service (RFS) to 12–14 months, directly addressing the market’s most acute pain point: hyperscalers and AI labs cannot wait 18–24 months for new capacity. The company’s ability to show active construction progress with secured supply chains and labor forces transforms sales conversations from speculative discussions into near-term commitment opportunities, as evidenced by immediate re-engagement from three hyperscalers after exclusivity expired at Pecos and Muskogee. This execution-driven model, validated by five operational CoreWeave sites generating over $350 million in annualized GAAP revenue, allows Core Scientific to monetize its development pipeline faster than pure-play developers who rely on lease signings before breaking ground, turning infrastructure readiness into a self-reinforcing demand accelerator.
The strategic shift to milestone-based exclusivity arrangements—replacing open-ended exclusivity with clear, time-bound milestones—eliminates a major historical drag on asset utilization while preserving negotiation leverage, enabling Core Scientific to return high-value sites like Pecos and Muskogee to market immediately when discussions stall, thereby avoiding costly opportunity costs from idle capacity. This approach, confirmed by management’s statement that they “will not keep high-value assets off the market longer than necessary,” directly addresses investor concerns about prolonged exclusivity periods stalling revenue recognition, and is already yielding results with three hyperscalers re-engaging within days of expiration. Combined with the company’s proven ability to scale sites to 1.5 gigawatts gross power at both Pecos and Muskogee through behind-the-meter gas pipelines, grid expansion, and the Polaris acquisition (440 megawatts), this flexible commercial strategy ensures that development progress never stalls due to contractual bottlenecks, allowing continuous capital deployment and value creation across the portfolio. The behind-the-meter infrastructure, while requiring air quality permitting, offers comparable power economics to grid-supplied power while delivering N+30% redundancy—critical for high-availability AI workloads—and reduces reliance on volatile utility interconnection timelines, a structural advantage in an era of grid congestion and permitting delays.
Core Scientific’s transition from Bitcoin mining to high-density colocation is not merely a wind-down but a strategic asset repurposing that unlocks embedded value in existing power infrastructure, with the company retaining only one or two mining sites by year-end while converting the majority of its Texas and Oklahoma campuses to HDC. This shift is de-risked by the fact that colocation revenue from 243 megawatts already covers operating costs and is expanding margins to 80%-85% cash gross profit—up from the original 75%-80% guidance—based on two years of operational experience with actual contractor deployment and headcount data, not theoretical models. The margin expansion is further supported by the lockbox structure in the $3.3 billion bond financing, which prioritizes operating expenses and debt service before allowing surplus cash flow to fund new projects, creating a self-funding growth cycle where CoreWeave-generated cash directly finances expansion at non-CoreWeave sites. Unlike peers reliant on volatile Bitcoin mining revenue or speculative development, Core Scientific now has a stable, contracted revenue base exceeding $350 million annually (straight-line over 12 years) that provides predictable cash flow to support its $2 billion 2026 capex plan, including $700 million for Hunt County and Polaris acquisitions, turning what was once a cost center into a scalable, high-margin growth engine.
Despite the optimistic narrative around behind-the-meter power solutions, Core Scientific faces significant, underappreciated execution risks in securing air quality permits for natural gas generation at Pecos and Muskogee, a process management acknowledges requires active participation with third-party operators and is already “far down the path” of—but not yet completed. The company’s reliance on behind-the-meter infrastructure to achieve its 1.5 gigawatt expansion targets introduces regulatory and timing uncertainties that could delay RFS beyond the 12–14 month window, particularly if Oklahoma or Texas environmental agencies impose stringent emissions controls or require costly retrofits for low-emission technology compliance, which would increase capital costs and erode the projected cost parity with grid-supplied power. This risk is exacerbated by the company’s admission that it has not conducted equivalent due diligence for behind-the-meter at the Hunt County site, suggesting a potential bottleneck in replicating the Pecos/Muskogee model across its broader portfolio, and could force reliance on slower utility interconnection processes, undermining the core thesis of accelerated time-to-market.
The $2 billion 2026 capital expenditure plan, while presented as disciplined, carries substantial risk of overruns and misallocation given the scale of concurrent projects—Hunt County acquisition, Polaris integration, Pecos/Muskogee expansion, Dalton Phase III, and Auburn—all advancing simultaneously with limited visibility into actual cost inflation for labor, equipment, and long-lead items like transformers and switchgear, which management admits are driving pricing firmness in new contracts. The company’s reliance on securing labor through large GCs with “leverage in the marketplace” does not guarantee cost containment, especially in a tight national construction market where wage pressures and subcontractor availability could trigger delays or premium payments, and the absence of explicit SG&A guidance despite a $30 million quarterly run-rate raises concerns about corporate overhead scaling disproportionately with growth. Furthermore, the lockbox structure, while innovative, creates a potential cash flow trap: if CoreWeave revenue falls short of projections due to tenant renegotiation or early termination (despite the 12-year term), the company could be forced to divert funds from new project development to cover operating expenses and debt service, stalling the very expansion it promises.
Core Scientific’s customer diversification narrative—emphasizing engagement with chip makers, AI labs, and Neo Cloud providers—overlooks a critical structural hurdle: these emerging segments often require additional credit support and lack the balance sheet strength of hyperscalers, making long-term, financeable commitments difficult to secure without costly guarantees or equity participation from Core Scientific. Management’s admission that they are “actively working with customers and financing partners on structures that can support long-term financeable commitments” implies that current customer conversations may not translate into near-term revenue, and the company’s focus on showing the *same sites* to all customer types risks diluting its sales effort without yielding signed contracts, particularly as hyperscalers remain the primary focus but are now evaluating multiple developers with proven scale. The company’s competitive advantage—five operational CoreWeave sites—may not be sufficient to win deals against entrenched players like Equinix or Digital Realty, which offer broader global footprints and established relationships with enterprise AI customers, and the absence of disclosed pricing trends or win rates in these emerging segments suggests the pipeline may be less concrete than implied, leaving revenue growth overly dependent on the CoreWeave contract’s 590 megawatt ramp, which itself faces execution risk if billable capacity falls short of the 450-megawatt summer target or 590-megawatt early 2027 goal due to unforeseen construction delays.
Despite the optimistic narrative around behind-the-meter power solutions, Core Scientific faces significant, underappreciated execution risks in securing air quality permits for natural gas generation at Pecos and Muskogee, a process management acknowledges requires active participation with third-party operators and is already “far down the path” of—but not yet completed. The company’s reliance on behind-the-meter infrastructure to achieve its 1.5 gigawatt expansion targets introduces regulatory and timing uncertainties that could delay RFS beyond the 12–14 month window, particularly if Oklahoma or Texas environmental agencies impose stringent emissions controls or require costly retrofits for low-emission technology compliance, which would increase capital costs and erode the projected cost parity with grid-supplied power. This risk is exacerbated by the company’s admission that it has not conducted equivalent due diligence for behind-the-meter at the Hunt County site, suggesting a potential bottleneck in replicating the Pecos/Muskogee model across its broader portfolio, and could force reliance on slower utility interconnection processes, undermining the core thesis of accelerated time-to-market.
The $2 billion 2026 capital expenditure plan, while presented as disciplined, carries substantial risk of overruns and misallocation given the scale of concurrent projects—Hunt County acquisition, Polaris integration, Pecos/Muskogee expansion, Dalton Phase III, and Auburn—all advancing simultaneously with limited visibility into actual cost inflation for labor, equipment, and long-lead items like transformers and switchgear, which management admits are driving pricing firmness in new contracts. The company’s reliance on securing labor through large GCs with “leverage in the marketplace” does not guarantee cost containment, especially in a tight national construction market where wage pressures and subcontractor availability could trigger delays or premium payments, and the absence of explicit SG&A guidance despite a $30 million quarterly run-rate raises concerns about corporate overhead scaling disproportionately with growth. Furthermore, the lockbox structure, while innovative, creates a potential cash flow trap: if CoreWeave revenue falls short of projections due to tenant renegotiation or early termination (despite the 12-year term), the company could be forced to divert funds from new project development to cover operating expenses and debt service, stalling the very expansion it promises.
Core Scientific’s customer diversification narrative—emphasizing engagement with chip makers, AI labs, and Neo Cloud providers—overlooks a critical structural hurdle: these emerging segments often require additional credit support and lack the balance sheet strength of hyperscalers, making long-term, financeable commitments difficult to secure without costly guarantees or equity participation from Core Scientific. Management’s admission that they are “actively working with customers and financing partners on structures that can support long-term financeable commitments” implies that current customer conversations may not translate into near-term revenue, and the company’s focus on showing the *same sites* to all customer types risks diluting its sales effort without yielding signed contracts, particularly as hyperscalers remain the primary focus but are now evaluating multiple developers with proven scale. The company’s competitive advantage—five operational CoreWeave sites—may not be sufficient to win deals against entrenched players like Equinix or Digital Realty, which offer broader global footprints and established relationships with enterprise AI customers, and the absence of disclosed pricing trends or win rates in these emerging segments suggests the pipeline may be less concrete than implied, leaving revenue growth overly dependent on the CoreWeave contract’s 590 megawatt ramp, which itself faces execution risk if billable capacity falls short of the 450-megawatt summer target or 590-megawatt early 2027 goal due to unforeseen construction delays.