Core Scientific Posts $232.9 Million Q3 2026 Revenue as Colocation Dominates

Core Scientific (NASDAQ: CORZ) reported $232.9 million in revenue for the third quarter of 2026, with colocation services — not bitcoin mining — as the main contributor, according to the company's quarterly disclosure. The figure marks a $68.7 million increase over the second quarter of 2026, when the firm booked $164.2 million in total revenue. Put plainly, the company added roughly the annual budget of a mid-sized hospital in a single quarter of growth.

The operational shift is now structural. In Q2 2026, colocation generated $136.7 million of revenue while self-mining brought in just $21.5 million — a gap of more than six to one. Core Scientific has not yet released a full Q3 breakdown by business line, and as of November 5, 2026, key profitability metrics such as Adjusted EBITDA remained unpublished. Revenue shows what came in; Adjusted EBITDA would show what the company kept after its enormous electricity bills. Until that number appears, the quarter's health rests on one figure alone.

What is driving the pivot?

Core Scientific increasingly functions as a landlord for AI hardware rather than a miner. Under its high-density colocation (HDC) model, customers such as CoreWeave and AMD bring their own servers and chips; Core Scientific supplies the buildings, power, cooling and connectivity. The economics differ fundamentally from mining. Self-mining income depends on token prices, network difficulty and shrinking block rewards — variables a miner cannot control. A colocation contract swaps that volatility for a tenant paying for capacity on fixed terms.

The scale is substantial. Core Scientific operates approximately 2.1 gigawatts of gross utility power across its US sites. Around 1.1 gigawatts of that capacity is already contracted to customers, and as of mid-2026 those contracts were linked to more than $24 billion in potential future revenue. That backlog equals roughly two hundred times the company's current quarterly colocation run rate — if, and only if, contracted power converts into billable capacity. Recent leases have covered hundreds of megawatts across multiple states.

The turnaround follows a Chapter 11 bankruptcy filing, from which the company emerged in January 2024. Fewer than three years later, it reports quarterly revenue in the hundreds of millions, largely from customers with no connection to bitcoin mining.

What should investors watch next?

Four unresolved questions define the outlook:

• Margins. Once published, Adjusted EBITDA will reveal whether colocation growth translates into actual profit rather than revenue alone.
• Backlog conversion. The $24 billion in potential future revenue is the core of the investment case; the pace at which 1.1 GW of contracted power becomes billable capacity will determine whether that figure is real or aspirational.
• Remaining capacity. With roughly 1 GW of gross power still uncontracted, the company retains room to sign additional tenants.
• Customer concentration. CoreWeave and AMD anchor the roster, but heavy dependence on a few large tenants means a single contract change could carry outsized consequences.

The concentration point deserves emphasis. The same client names that validate the model also concentrate its risk. Core Scientific has not disclosed the terms, durations or renewal conditions of its major leases, and those omissions matter more than headline revenue.

The next disclosure window — when full Q3 segment data and Adjusted EBITDA arrive — will test whether the colocation thesis holds up under scrutiny. For now, the market is being asked to trust a growth story whose profitability half remains unpublished.