Nasdaq-listed Cipher Digital reported its first high-performance computing (HPC) lease revenue in the third quarter of 2026 as it expands its AI infrastructure business beyond bitcoin mining.
Cipher reported $27 million in quarterly revenue and a loss of $10 million in adjusted EBITDA, including $5.9 million in newly recognized HPC leasing revenue. Analysts had estimated around $38 million in revenue.
The company extended its Barber Lake agreement with a leading AI lab by 10 years, bringing contracted revenue at the 300 MW site to approximately $9.1 billion, and said its development pipeline has grown to 5.9 GW, including plans for up to 2.5 GW of on-site power generation.
VanEck analyst Matthew Sigel noted the company is still expecting 2GW of grid power in 2028-2029, while the 2030 target for grid dropped from 2.1GW to 1.1GW from 2.1GW.
“Cipher was built for this market, and this quarter we took important steps to strengthen our position for the long term,” CEO Tyler Page said in a release. “The decisions we made early on about who we partner with, how we structure our projects, how we navigate a dynamic regulatory environment and how we engage with our communities were made for exactly these conditions."
Demand for compute is growing, Page said, and Cipher's relationships with leading hyperscalers and AI labs are deepening. "We are adding new capacity to our pipeline, investing earlier in every project and advancing design work before committing to new leases, so we are well positioned for our next phase of growth," he added.
On the company's earnings conference call, Sigel said management noted Odessa came out of the 2027 capacity count as it's still mining bitcoin, but HPC talks are ongoing.
"They described it as capital-light, with no Cipher equity and a return for unlocking tenant interest using the long lead equipment and PPA term," Sigel wrote in a post on X.
Cipher's shares (CIFR) dropped more than 5% after Monday's market open, but recovered to around unchanged at publication time. The stock is down 55% in the year-to-date period.