> ## Content Index
> Fetch the complete content index at: https://www.metatalks.ai/llms.txt
> Use this file to discover other available public pages before exploring further.

# Bitcoin miners are building AI capacity far faster than the AI work pays for it
- URL: https://www.metatalks.ai/bitcoin-miners-six-month-capital-spending-already-tops-their-whole-2025-total/
- Published: 2026-08-24T16:41:00.000Z
- Updated: 2026-08-24T16:40:59.000Z
- Author: Al
- Tags: News, AI Infrastructure, mining, AI Economy, #newswire

**Nine of them break the AI revenue out separately, and it comes to one dollar for every fifteen of capital spending.**

Twelve listed bitcoin miners tracked in [an industry dataset](https://www.minerweekly.com/p/307-billion-miners-and-ai-peers-already?ref=metatalks.ai) laid out a net $6.87 billion of cash for capital assets over the first six months of 2026, more than the $6.50 billion the same group spent across all of 2025\. IREN enters that total only through its quarter ended in March, because no cash-flow statement covering the June period was available.

[Data centers and AI work](https://cointelegraph.com/news/bitcoin-miners-ai-hpc-capex-revenue-2026?ref=metatalks.ai) have been pitched to miners as a way to broaden the business while conditions in mining stay tough, and the switch is being paid for up front.

First-half capital outlays came to $1.61 billion at TeraWulf, $1.58 billion at Applied Digital, $1.18 billion at Core Scientific and $911.5 million at Cipher.

That revenue — booked explicitly as HPC, AI cloud and colocation work by the nine miners that broke it out in both quarters — reached $205.8 million in the second quarter of 2026, 52% above the first quarter's $135.4 million. Most of the dollar-value increase came from Core Scientific, whose colocation revenue rose to $136.7 million from $77.5 million. TeraWulf's HPC leases brought in $31.9 million, up from $21 million, and Bitdeer's AI Cloud $14 million, against $3.7 million before. Narrowed to those same nine, capital spending over the half year came to $5.11 billion against $341.2 million of AI revenue, roughly fifteen to one.

Capital spending runs ahead of the revenue it is meant to produce for structural reasons. Electricity contracts and land already in hand can give miners an early edge, but turning those holdings into capacity fit for AI takes substations, buildings, cooling, networking gear and, under some business models, GPUs — and most of that money goes out before any matching revenue can be booked.

Setting one period's revenue against capital spending is not a standard gauge of how profitable a project is. Capital outlays create assets meant to earn across several years, while a single quarter's revenue reflects only capacity that has been handed over, signed off and put to work, and customer prepayments can bankroll construction without showing up as revenue right away. Even so, that gap holds the chief danger of the move into AI: the money is committed at once and is mostly impossible to recover, while converting it into revenue depends on how fast projects are built, when grid hookups arrive, whether customers accept the capacity and whether demand holds.

For the rest of 2026, then, the open question is not whether AI-related revenue keeps rising — company filings make plain that it is — but whether it can begin closing the distance on the record sums already put up to generate it.