Bitdeer AI, part of Bitdeer Technologies Group, said on 8 October 2026 that it has entered into a 10-year data center services agreement for A901, a 67MW AI Cloud data center campus in Malaysia. Delivery of capacity is targeted to begin in the first quarter of 2027, with the full 67MW targeted for the second quarter of 2027.
The release also states that no offtake commitments have been entered into with respect to A901 capacity as of its date, and that its active pipeline figure of more than $10 billion is a management estimate, not backlog.
Business Pill · BACKLOG
A short explainer of backlog: revenue that is signed but not yet earned. It teaches the general idea only and says nothing about any company in this story.
The key insight: As we read it, the release separates three things that are easy to blur: capacity Bitdeer has secured, capacity under customer offtake, and opportunities it is discussing. By its own account, A901 is in the first group but not yet the second.
According to the release, A901 is Bitdeer AI’s third Malaysian hub alongside Cyberjaya and Johor, and will be engineered for liquid-cooled, rack-scale NVIDIA GB300 NVL72 systems. All megawatt figures refer to critical IT load, it says.
With A901, Bitdeer AI says its secured AI Cloud data center capacity, meaning capacity owned or under an executed data center services agreement, totals approximately 273.5MW across Malaysia, Norway and the United States. That is approximately 78% of its target of up to 350MW by the first quarter of 2028.
On our arithmetic, that implies about 206.5MW secured before A901 and about 76.5MW still to secure against the target.
Bitdeer AI’s secured AI Cloud data center capacity in MW of critical IT load, per its release of 8 October 2026: about 273.5MW with A901 against a target of up to 350MW; the 206.5MW before A901 is our arithmetic.Per Bitdeer AI’s release of 8 October 2026: no offtake commitments for A901 capacity as of its date; a 67MW campus under a 10-year agreement, with delivery targeted from the first quarter of 2027; a pipeline estimate above $10 billion that the release says is not backlog.
The Pipeline and Its Caveats
The release describes demand for uncontracted capacity as strong and estimates the active pipeline for AI Cloud capacity at more than $10 billion, which it expects to keep increasing. Contract value per megawatt varies with service mix, contract duration and site, it says.
Its forward-looking statements define that pipeline as the estimated aggregate potential contract value of opportunities under commercial discussion, and say it is not backlog, revenue, an executed contract or a binding commitment.
“Demand for liquid-cooled, rack-scale AI capacity in Southeast Asia is running ahead of what the market can deliver, and customers are asking for it earlier and in larger blocks,” said Michael G. Potter, Chief Financial Officer of Bitdeer Technologies Group.
A Withdrawn Release
On 8 October, GlobeNewswire also published a notice saying it had been advised by Bitdeer Technologies Group that readers should disregard a release issued earlier that day, titled “Bitdeer AI Secures 60MW A103 Facility at Its Existing Cyberjaya, Malaysia Campus; Reaches Approximately 95% of Its Up-to-350MW Target Capacity.” The notice we read gives no reason.
The Structural Read
Secured means owned or under agreement. The release defines secured capacity as capacity owned or under an executed data center services agreement.
The pipeline is an estimate. Bitdeer describes it as the aggregate potential contract value of opportunities under commercial discussion.
Timing is targeted for 2027. Capacity delivery is targeted to begin in the first quarter of 2027, with the full 67MW in the second quarter.
Michael G. Potter, CFO, Bitdeer Technologies Group, 8 October 2026
“A901 adds 67MW in a third Malaysian location, which widens the set of sites and delivery dates we can put in front of a customer.”
Three Implications
ABOUT 78% OF TARGET Secured capacity of approximately 273.5MW is about 78% of the up-to-350MW target, per the release.
NO OFFTAKE YET The release says no offtake commitments for A901 had been entered into as of its date.
TO BE ENGINEERED FOR GB300 NVL72 A901 will be engineered for liquid-cooled, rack-scale NVIDIA GB300 NVL72 systems, Bitdeer says.
The framework’s starting point: “The structural bottleneck isn’t computing, it’s electrons.”
As we read it, Bitdeer’s release measures progress in megawatts of critical IT load secured, and for A901 no customer offtake had been signed as of the release date, according to the release.
What Is Not Established
We read Bitdeer AI’s release and the GlobeNewswire notice in full. The release we read does not name the data center provider or state the agreement’s value. We did not contact Bitdeer.
Business Engineer Framework
The State of AI Data Centers
A Business Engineer framework on power, space and capital as the constraints on AI infrastructure.
Bitdeer AI says its 10-year agreement for the 67MW A901 campus takes its secured AI Cloud capacity to approximately 273.5MW, about 78% of its target of up to 350MW by the first quarter of 2028. The release also states that no offtake commitments for A901 capacity had been entered into as of its date, and that its pipeline estimate of more than $10 billion is not backlog.
A note on sourcing. We read Bitdeer AI’s release of 8 October 2026 and the GlobeNewswire notice in full; the capacity before A901 and the remaining gap are our arithmetic. We did not contact Bitdeer. Nothing here is a forecast, and nothing here is financial or investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
