Core Scientific built its business around Bitcoin mining. Now, the company’s latest financial results show that artificial intelligence infrastructure has become its dominant source of revenue. The shift raises a question that reaches well beyond one company: what happens when operating data centers becomes more attractive than mining cryptocurrency?
The Core Scientific AI pivot provides a different perspective on the relationship between two industries competing for electricity and computing infrastructure. While investors have debated Bitcoin and AI market rotation, this development concerns the businesses supplying computing power. For some mining operators, the more valuable asset may no longer be the Bitcoin they produce, but the electricity and facilities they control.
Core Scientific AI Pivot Is Already Visible in Revenue
Core Scientific’s November 5 earnings release revealed how dramatically its business has changed.
The Nasdaq-listed company generated $232.9 million in third-quarter revenue, with $215.4 million coming from high-density colocation services. That means approximately 92.5% of quarterly revenue came from its data-center business rather than cryptocurrency mining.
By comparison, colocation revenue was just $15 million during the third quarter of 2025.
The company’s third-quarter financial results also showed adjusted EBITDA reaching $133.6 million, compared with $6.5 million a year earlier.
Reported net income reached approximately $1.2 billion, although that figure requires caution. Much of the profit resulted from a noncash change in the fair value of warrants, rather than ordinary data-center operations.
The distinction matters because investors evaluating the transition should focus on recurring operating revenue, construction spending and cash generation rather than treating the headline profit as a straightforward measure of business performance.
Bitcoin Mining Is Losing Its Place in the Revenue Mix
The transformation becomes even clearer when examining the first nine months of 2026 against the same period in 2025.
| Revenue Segment | First Nine Months of 2025 | First Nine Months of 2026 |
|---|---|---|
| AI and high-density colocation | 14% | 84% |
| Bitcoin self-mining | 78% | 12% |
| Hosted cryptocurrency mining | 8% | 4% |
| Total revenue | $239.3 million | $512.3 million |
The figures reveal a fundamental business change.
Core Scientific was primarily a Bitcoin mining company in 2025. By September 2026, it had become predominantly a provider of data-center infrastructure for customers running demanding computing workloads.
Bitcoin mining profitability depends on cryptocurrency prices, mining difficulty, electricity costs and equipment efficiency. AI infrastructure hosting operates under different economics, often involving longer-term customer contracts for computing capacity, electricity and cooling.
Those contracts may provide more predictable revenue, although they also introduce construction, financing and customer-concentration risks.
The important distinction is that revenue predictability does not guarantee profitability. Building and operating high-density data centers requires considerable capital, and signed agreements do not immediately become recognized revenue.
The Company Is Selling Bitcoin to Finance Its Transition
Core Scientific is also reducing its direct exposure to Bitcoin ownership.
Its September 2026 quarterly filing reveals that the company mined approximately 869 BTC during the first nine months of the year.
Over that same period, it sold 3,374 BTC, generating approximately $286.8 million to support capital expenditures and other cash requirements.
By September 30, its Bitcoin holdings had fallen to 32 BTC, compared with 2,537 BTC at the end of December 2025.
Those figures are particularly revealing. Core Scientific sold considerably more Bitcoin than it produced during the period, drawing down its existing holdings as its infrastructure business expanded.
The decision should not automatically be interpreted as a prediction that Bitcoin prices will decline. The company’s stated strategy is to redirect resources toward its data-center operations.
Nevertheless, the transactions demonstrate that Bitcoin is becoming less central to the company’s balance sheet and operating model.
Core Scientific also recorded a $266.5 million impairment charge on mining-related assets during the first quarter of 2026, reflecting deteriorating mining economics and declining equipment values.
Together, these developments show why management is changing its priorities.
AI Infrastructure Offers Opportunity but Creates New Financial Risks
The attraction of AI infrastructure becomes clearer when examining Core Scientific’s contracted capacity.
As of September 30, the company had approximately 1,120 megawatts of contracted customer power capacity, with 485 megawatts already generating billable revenue.
That means roughly 43% of contracted capacity was billing customers at the reporting date, leaving substantial infrastructure still to be delivered.
The company also announced a July agreement involving AMD and a cloud infrastructure counterparty, covering approximately 530 megawatts across five sites.
Revenue from that arrangement is expected to begin in phases during 2027 as facilities are completed and delivered.
However, growth comes at a considerable cost. Core Scientific reported approximately $1.5 billion in capital expenditures during the first nine months of 2026.
The company must continue financing construction, obtaining equipment, completing utility connections and bringing facilities online.
There is also customer concentration risk. Its existing colocation revenue remains heavily dependent on CoreWeave, making successful diversification particularly important.
These are different risks from mining Bitcoin, but they are not necessarily smaller.
What This Shift Means for the Bitcoin Mining Industry
Core Scientific’s transition offers a useful way to evaluate other publicly traded mining companies.
The old analysis focused heavily on Bitcoin production, mining difficulty, electricity prices and the number of coins held in corporate treasuries.
Those metrics still matter for companies whose primary business remains mining.
But operators converting facilities into AI data centers require additional scrutiny. Investors must examine contracted capacity, megawatts actually generating revenue, construction costs, debt obligations and the financial strength of major customers.
A company announcing an AI hosting agreement has not necessarily completed the infrastructure or started earning revenue from it.
That distinction will become increasingly relevant as mining companies compete for customers seeking access to electricity, data-center space and advanced cooling systems.
The transition also does not mean Bitcoin mining itself is disappearing. The Bitcoin network continues to rely on miners, and individual companies can reduce participation without changing the network’s underlying rules.
The more consequential question is whether other operators will reach similar conclusions about their own facilities.
Core Scientific’s latest results establish a significant benchmark: a business that once depended overwhelmingly on Bitcoin mining now generates most of its revenue from another industry.
The next test is execution. Delivering contracted capacity, controlling construction costs and reducing customer concentration will determine whether the transition produces sustainable returns.
The Core Scientific AI pivot illustrates an important development in cryptocurrency infrastructure. As electricity and computing capacity become increasingly valuable to AI companies, some Bitcoin miners may find that their strongest commercial opportunity lies outside cryptocurrency itself.
