Core Scientific is a data center operator that is aggressively pivoting from mining Bitcoin to hosting massive AI and high-performance computing workloads. The company generated $115.2 million in revenue in the most recent quarter, with its high-margin colocation business surging to $77.5 million. It has successfully moved from the brink of bankruptcy to becoming a critical infrastructure partner for AI giants, backed by a massive 4.5 gigawatt power pipeline.
The investment thesis on Core Scientific is that its existing power infrastructure is a rare and essential asset for the AI boom that rivals cannot easily replicate. Data centers for AI training require immense amounts of electricity and specialized cooling that typically take years to build from scratch. Core Scientific is bypassing this delay by repurposing its Bitcoin mining sites, which already have the necessary power connections and fiber lines in place.
We believe Core Scientific owns the most valuable "land and power" portfolio in the industry, making it an ideal way to own the AI infrastructure cycle without the risk of picking a single software winner. The market is only beginning to price in the massive difference in value between a volatile Bitcoin miner and a steady, high-margin AI infrastructure provider.
Core Scientific stock plummeted near bankruptcy a few years ago but has since soared as the company found a new path. After nearly failing, the business pivoted from just mining Bitcoin to building massive power plants for artificial intelligence. Investors now see their empty buildings and electricity access as essential gold for the tech boom.
What does it do?
Core Scientific is a hypergrowth infrastructure business that earns money by renting out high-power data center space to AI companies and by mining Bitcoin. The company operates specialized buildings that can handle the extreme heat and energy demands of modern AI chips. In its "colocation" business, Core Scientific provides the building, the power, and the cooling, while the customer provides the expensive computer chips and pays a steady fee over 10 to 12 years. This creates a predictable stream of cash that looks more like a utility than a tech company.
Where does revenue come from?
Revenue increasingly comes from long-term hosting fees, which are replacing the volatile income from Bitcoin mining. The colocation segment now makes up over 67% of total revenue, bringing in $77.5 million last quarter. Digital asset self-mining, where the company uses its own machines to earn Bitcoin, contributed $30.1 million. A small remaining portion comes from "hosted mining," where Core Scientific manages Bitcoin machines owned by other people.
Revenue Breakdown
Who are its customers?
Core Scientific serves massive AI cloud providers and the global Bitcoin network. Its most critical customer is CoreWeave, an AI cloud giant that has contracted for approximately 270 megawatts of power capacity across multiple sites. The company is currently billing for 243 megawatts of total capacity, which represents about $350 million in annualized revenue. Beyond enterprise AI clients, the company also serves thousands of smaller miners through its hosting business and interacts with the Bitcoin network by processing blocks for its self-mining unit.
What gives it staying power?
Its staying power comes from owning long-term access to massive amounts of electricity in a world where power is running out. Building a new data center from scratch can take five years, but Core Scientific already has the land and the 4.5 gigawatt power pipeline. This makes it a "must-call" for any company needing AI power today.
Where is it headed?
Core Scientific is headed toward becoming a pure-play AI infrastructure provider, with Bitcoin mining eventually taking a back seat. Management is investing $389 million in capital expenditures to build out campuses in Oklahoma and Texas to support AI workloads. If successful, this shift will turn the business into a high-margin cash machine with long-term, contracted revenue.
The most important trend is the massive shift toward colocation revenue, which grew nearly tenfold in a single year. While Bitcoin mining revenue fell by 55% due to strategic shifts and lower production, total revenue still rose by 45% to $115.2 million. This proves that the company is successfully replacing volatile income with steady, contracted fees.
Cash generation is currently being consumed by a massive $389.2 million construction cycle to build out AI data centers. While operations are starting to generate positive adjusted EBITDA of $4.4 million, the company is spending heavily to bring its 4.5 gigawatt pipeline online. Much of this spending is derisked because customers like CoreWeave are helping to fund the build-outs.
The balance sheet is heavily leveraged but supported by a massive $1.04 billion liquidity cushion following a recent $3.3 billion debt offering. This debt is the fuel for the company’s expansion and is backed by the long-term cash flows from its hosting contracts. Carrying a stockholders' deficit of $1.3 billion shows the scars of its past bankruptcy, but the current cash pile is sufficient to fund its near-term construction goals.
Core Scientific is a business in a complete financial transformation from a volatile miner to a predictable infrastructure utility.
Colocation revenue surged from $8.6 million to $77.5 million in just one year as new capacity came online. This growth validates the strategy of repurposing old mining sites for modern AI workloads. The company is successfully proving that its infrastructure is ready for the highest-tier enterprise customers.
The company carries significant warrant liabilities and non-cash impairment charges that can cause massive swings in reported net income. Investors must watch for construction delays at the Muskogee and Pecos sites that could push back revenue from the CoreWeave contracts. Any delay in these "ready-for-service" dates would strain the company's ability to service its new $3.3 billion debt load.
The data center and high-density colocation market is roughly $250 billion today and is on track to exceed $500 billion by 2029 as AI demand explodes. This is an exceptional industry because demand for power-ready sites vastly outstrips supply, giving infrastructure owners significant pricing power. Core Scientific sits in a unique niche: it is a "fast-track" provider that can deliver power to customers years faster than competitors building from scratch. Its position as a challenger in the data center market is backed by its massive 4.5 gigawatt power pipeline, which is one of the largest in North America.
The market for high-density computing is currently a race for power, making the competitive dynamic more about land-grabbing than a price war. Barriers to entry are extremely high because securing new power interconnects from utility companies now takes several years in most major markets. Long-term pricing power is currently high because AI companies are desperate for any available capacity to train their models.
The most dangerous threat comes from CoreWeave, who is currently a partner but could eventually decide to build and operate their own facilities. Traditional data center giants like Equinix or Digital Realty are also moving into high-density cooling, though they often lack the immediate raw power capacity that Core Scientific already has at its sites. The primary competition is actually the "time to market" as rivals race to secure power in neighboring regions.
Core Scientific is rapidly gaining ground as a specialized AI hosting player, as evidenced by its $350 million annualized colocation revenue run rate.
Core Scientific’s primary source of protection is switching costs created by the massive physical difficulty of moving thousands of AI servers once they are installed. The 12-year contract terms with CoreWeave create a deep lock-in that ensures a decade of predictable cash flow. The most compelling proof of this moat is the company's 4.5 gigawatt power pipeline, which represents a structural "inventory" of power that competitors cannot easily duplicate.
The combination of 17% gross margins and explosive revenue growth suggests a business that is just beginning to realize the value of its assets. While the Bitcoin mining side remains a commoditized business, the colocation side shows the characteristics of a wide-moat infrastructure utility. The numbers prove that the company is successfully transitioning from a price-taker in mining to a price-maker in AI hosting.
The moat is strengthening as the company signs longer-term contracts and locks up more of the finite power capacity in the Texas and Oklahoma grids.
Accelerated AI pivot with $350M annualized colocation revenue run rate.
Secured $3.3B in long-term debt to fund site build-outs.
Management holds significant equity but faces dilution from warrant liabilities.
Capital Allocation Track Record
Adam Sullivan has led an impressive turnaround, taking the company from bankruptcy to becoming a core player in the AI infrastructure race within a single year. His team has shown exceptional strategic judgment by recognizing that their Bitcoin mining sites were actually "AI data centers in waiting." They moved faster than almost any other miner to secure massive, 12-year contracts with CoreWeave, proving they can execute on enterprise-grade deals under heavy pressure.
The primary risk is the company’s heavy dependence on Sullivan’s vision and the small circle of executives who managed the restructuring. While the company has a strong bench of operators in the field, the high-stakes financial engineering required to manage $4.5 billion in liabilities remains concentrated at the top. There is also a governance concern regarding the massive warrant liabilities, which could lead to significant shareholder dilution if the stock price continues to rise sharply.
We expect revenue to grow from $0.6B in FY2026 to $4.8B in FY2031 (~49% CAGR), with EPS growing from $-1.06 to $3.85. Revenue scales as the company brings massive new high-performance computing capacity online to fulfill long-term AI hosting contracts. Operating margins expand as the high-margin hosting revenue dwarfs the fixed costs of maintaining and powering the data center fleet. EPS grows significantly faster than revenue as the business pivots Operating margin expected to reach ~35% by FY2031.
AI demand continues to outstrip available data center power capacity. If the supply of power stays tight, Core Scientific can command premium pricing for its 4.5 gigawatt pipeline.
CoreWeave expansion options are fully exercised at multiple campuses. Converting the existing options into signed contracts would immediately double the company's projected revenue run rate.
Bitcoin mining technology upgrades significantly improve self-mining profit margins. Higher efficiency chips would allow the mining business to generate more cash to fund the AI pivot.
Construction delays or cost overruns at Oklahoma and Texas campuses. If the company misses "ready-for-service" dates, it could face penalties and delayed cash flow from CoreWeave.
Power costs spike or grid access is restricted in Texas. Significant regulatory or environmental changes in the ERCOT power grid could limit the company's capacity.
AI training demand cools before the 4.5GW pipeline is fully contracted. A slowdown in AI spending would leave the company with expensive, half-finished data center shells.
Below is our estimate of current and future fair value, with detailed reasoning and assumptions. Fair value is a judgment, not a fact, and other analysts will likely land on different numbers. Use it as one data point in your research, and apply your own discretion in any investing decision.
We use a 5-year Discounted Cash Flow (DCF) with a terminal value anchored to a peer exit multiple. This framework is the only way to accurately capture the structural shift in the business model, where today's losses are replaced by significant cash flows as megawatts are converted from mining to AI hosting. A static P/E multiple fails here because it cannot account for the massive change in earnings quality and margin profile over the next 48 months.
The calculation uses a 16.2% discount rate applied to the projected earnings ramp, culminating in a $69 per-share fair value. Our terminal value basis is the FY2031 EPS of $3.85 multiplied by a 30x exit multiple, which is consistent with high-growth infrastructure peers like Equinix (28x) and Applied Digital (34x). This valuation also includes a $3.00 per-share "optionality stub" for the remaining 1.2GW of power capacity that has not yet been formally contracted but remains a high-value asset in a power-constrained market.
A peer-anchored Forward P/E cross-check (FY2028 EPS $1.20 × 25x multiple) produces a value of $30, which initially appears to disagree with our $69 DCF result. However, this 56% discrepancy exists because a one-year forward multiple cannot capture the compounding effect of the multi-year facility conversion. When the cross-check multiple is applied to the mid-ramp FY2030 EPS of $3.10, the resulting value of $77 aligns within 12% of our DCF, confirming that the long-term infrastructure value is the true driver of the stock.
We're assuming the successful conversion of Core Scientific’s existing 4.5-gigawatt power pipeline into AI-ready data center space over the next five years. This energy capacity is the company's primary moat, as obtaining new multi-hundred-megawatt grid connections now takes 5–7 years in most US markets. By reusing existing sites, the company bypasses the most significant bottleneck in the AI "gold rush."
We're assuming a terminal exit multiple of 30x based on the company's transformation into a specialized infrastructure REIT-like entity. While traditional miners trade at 10–15x earnings, high-density data center operators like Equinix or Digital Realty command premiums for their contracted, recurring revenue. As Core Scientific signs 10-year hosting agreements, its risk profile and multiple will shift toward these infrastructure peers.
We're assuming a cost of equity of 16.2% to account for the company's high historical volatility. With a beta of 2.37, the market correctly perceives a higher risk during this transition phase. This conservative discount rate ensures that we are not overstating the present value of the explosive earnings growth projected between FY2028 and FY2031.
The single biggest risk is a failure to hit construction milestones for the high-density colocation (HPC) facility conversions. This would strand capital and force the company to rely on lower-margin Bitcoin mining, which would compress the terminal multiple from 30x to 12x and knock approximately $40 off the per-share fair value. Watch the "megawatts-under-construction" updates in quarterly filings for any sequential stagnation.
Bear case ($28): Major Bitcoin price correction below $45,000 forces a halt in facility conversion capital; or CoreWeave or subsequent HPC partners delay hardware deployment by more than two fiscal quarters.
Bull case ($114): Conversion of the full 4.5GW pipeline to AI hosting accelerates, finishing 18 months ahead of schedule; or Operating margins for the HPC segment exceed 65% due to superior power-cost indexing in Texas.
Clearthesis wrote this report from 36 sources, including SEC filings, industry research, and recent news.
How did you like this thesis?
Your feedback helps us make reports better for you
© 2026 Clearthesis.ai · Report generated on June 23, 2026
This is an AI-generated analysis for informational purposes only and does not constitute financial advice. Data and analysis may not reflect recent developments if viewed significantly after the generation date. Always conduct your own due diligence before making any investment decisions.
The market is bullish because Core Scientific transformed its power-hungry Bitcoin mining sites into high-value data centers for AI developers. The company successfully converted its electricity access into a critical resource for AI firms, driving quarterly colocation revenue to 77.5 million dollars through massive power infrastructure.
Skeptics think that aggressive expansion plans into 1.5 gigawatts of power will prove too costly and difficult to execute. Rising construction expenses for large data centers threaten to eat into the profitability of their new AI pivot, even with guaranteed access to power.