Skip to content
World clockEU--:--UK--:--USA--:--CN--:--PLDEFRIT中文EN

portal about AI and technologyevents · analysis · interviews · technical background

Search
LIVE
›

Bitdeer secures 206.5 MW of AI cloud capacity as providers race to add regions

Bitdeer has secured about 206.5 MW of AI Cloud data-centre capacity in Malaysia, Norway and the United States, roughly 59% of its up-to-350 MW target by the first quarter of 2028, its investor relations head told TNGlobal on 4 November.

TechnologyNewsGrace OkonkwoPublished: 4 November 20264 min readSources 28
Bitdeer secures 206.5 MW of AI cloud capacity as providers race to add regions

Bitdeer puts a number on its AI cloud pivot

Pretesh Dahya, Bitdeer's head of investor relations, gave the figures to TNGlobal in an interview published on 4 November.

For enterprise buyers, timing matters more than the number. AWS capacity is largely sold out through 2027. Every megawatt a neocloud or converted miner brings online is capacity a buyer cannot rent elsewhere. It is still unknown how much of that space will be delivered on schedule, and at what price.

The hard part is not the GPUs. Bitdeer's disclosure shows how much of the work is grid infrastructure. It has contracted 121 IT MW at its Tydal campus in Norway under a 16-year AI/HPC co-location agreement. It owns or operates about 255 acres and 742 MW of existing and pipeline power capacity in Milam County, Texas. Dahya said the reusable assets are also the slowest to secure: land, large-load interconnections, transmission and substations. "Conversion is therefore not simply a matter of replacing mining rigs with GPUs," he said.

The company says it is not managing toward a fixed split between mining and AI. Capacity goes where contracted demand and returns are, it says.

The capacity crunch is already visible

Spending figures explain why providers keep building. Worldwide cloud infrastructure revenue reached $168 billion in the third quarter of 2026, up 55% from a year earlier, according to Synergy Research Group. The market doubled in eight quarters. AI-specific services now account for roughly one-third of total revenue. Synergy's John Dinsdale told Data Center Knowledge that much of the new capacity "is specifically targeting more valuable, higher-priced AI services". That mix shift helps explain how revenue can outrun the physical capacity behind it.

CBRE counted 7,481 MW under construction across eight primary North American markets in the first half of 2026, up 24.8% from a year earlier. More than 80% of that space was preleased, against 74.3% a year earlier. Power constraints kept pushing completion schedules out. Customers compete for capacity that has not been delivered.

The top three are not the fastest movers. Amazon held 26% of worldwide cloud infrastructure revenue, Microsoft 19% and Google 16%, Synergy reported. Oracle and neoclouds including CoreWeave, Crusoe and Nebius were among the fastest growers outside the top three. Synergy said neocloud expansion is eroding the share of some established providers.

Regions and dollars follow demand

Alibaba Cloud opened its first public cloud region in the Netherlands on 29 October, with two data centres in Amsterdam. Amsterdam is its fourth European region after Frankfurt, London and Paris. The region sits inside a previously announced $53 billion investment in AI and cloud infrastructure.

The company now runs 109 availability zones across 32 regions. It plans a first cloud region in Finland and further expansion in Germany and France over the next 12 months. Feifei Li, chief technology officer and president of international business at Alibaba Cloud, said the region gives customers "in-region infrastructure, performance and data residency". Data residency rules are the practical reason European buyers care about where a region sits, not just how many exist.

Hyperscalers accounted for 59% of North American data centre tenant demand, with 28 GW of owner-occupied capacity under construction, according to JLL's August midyear report.

Policy and credit risk shadow the buildout

Nvidia has told some cloud customers to take on less business from Chinese companies, The Information reported, citing three people familiar with the matter. The guidance is not mandatory. It applies to future contracts rather than existing deals. Nvidia also asks customers to disclose who is using the chips. The Financial Times reported Nvidia communicated the instruction verbally through sales reps and regional teams. The gap is remote access: US export rules restrict shipments, but generally not renting compute abroad.

Regulators had already been looking at the practice. US authorities looked into Chinese AI companies reaching such chips through foreign data centres after allegations involving Moonshot AI, the developer of the Kimi chatbot.

Nvidia's three largest customers generate 44% of revenue, according to an AFP summary cited by TradingView. Its five-year credit-default-swap spread rose to 127-129 basis points, up from 39 at the end of 2025 and 84 on 1 September, according to Benzinga.

Nvidia carries more of that risk itself. It has guaranteed land, power and building-shell obligations tied to data-centre leases for select AI cloud partners if those partners default. It also agreed to buy computing capacity it cannot sell to others through April 2032.

What to watch next

Nvidia reports quarterly results on 17 November. The report is expected to update demand from major cloud customers. Bitdeer's 350 MW AI Cloud target runs to the first quarter of 2028.

Comments 0

Sources

28
  1. 01Cloud Revenue Hits $168B as AI Capacity ExpandsEN
  2. 02Weekly Recap: Meta equity investment and Oracle Cloud Vera Rubin systemsEN
  3. 03Alibaba Cloud Launches New Cloud Region in AmsterdamEN
  4. 04Alibaba Cloud Launches First Public Cloud Region in the NetherlandsEN
  5. 05Alibaba Cloud Opens Amsterdam Region with Two Data CentresEN
  6. 06Nvidia pushes cloud customers to cut back on China businessEN
  7. 07Bitdeer on building AI/HPC infrastructure across three regionsEN
  8. 08AWS Capacity Largely Sold out through 2027, with On-Demand Headroom ReservedEN
  9. 09Waaree doubles Texas solar module manufacturing capacity to 3.2 GWEN
  10. 10Congress Is Cutting Off Transportation Investments that Help Families Save MoneyEN
  11. 11The ETS Investment Engine: Scaling Europe’s Strategic E-fuel Market for Shipping & AviationEN
  12. 12Google: LLMjacking attacks rack up $1 million in cloud AI chargesEN
  13. 13 Putin demands Russian data centers supply capacity to the state in wake of Ukraine drone attacks EN
  14. 14OpenAI deepens Abu Dhabi foothold as Sam Altman inks deals spanning energy, government and investmentEN
  15. 15Google adds new $250,000 Cloud VRP reward for Root-in-Cloud bugsEN
  16. 16DarkBlinders Compromises Israeli Target and Iraqi Kurdish Government Cloud EnvironmentEN
  17. 17Künstliche Intelligenz: Riskante Wette: KI‑Investitionen übersteigen eine Billion DollarEN
  18. 18Xiaomi follows Li Auto with planned Sunwoda EVB investment to deepen battery tiesEN
  19. 19Big Oil Cashes in on Trump Campaign Investment, Doubles Earnings Compared to Last YearEN
  20. 20Shopify Overexposed Scans GKE Ingresses and Cloud Run for ExposureEN
  21. 21Cloudflare’s Post-Quantum GeoKey Manager Enters Closed Beta without Attribute-Based EncryptionEN
  22. 22Google Cloud Day Turkey 2026: $2 Billion Investment ProgressesEN
  23. 23Cloudflare Reports 18.2 Million Network-Layer DDoS Attacks in Q3, Up 38%EN
  24. 2426,400 Valid Cloud and AI Credentials Found Exposed in Public Gitee RepositoriesEN
  25. 25Oracle Targets $225B Revenue by 2030 as AI Cloud Build-Out Accelerates - TradingViewEN
  26. 26Media rights squeeze, US investment and AI tech: The sports business trends from Sportel 2026EN
  27. 27Could new US investment curbs complicate China’s biotech licensing surge?EN
  28. 28How Ultrafast cut incident investigation from hours to minutes at OpenAIEN

All figures and quotations in this text come from the sources listed below.

Content prepared by the editorial team with AI assistance.

Grace Okonkwo

Grace Okonkwo

AI, models and technology

Grace Okonkwo covers AI, models and technology for FLASH24, working from primary sources such as model cards, API documentation and benchmark papers rather than vendor summaries. She checks training data provenance, evaluation conditions and reported scores against the underlying datasets before any figure reaches print. She interviews researchers and engineers directly, tracks release calendars from major labs, and compares successive model versions on the same tests. Her own self-hosting, home-network and documentation-reading habits feed straight into that desk, since she tests tools on her own hardware first. She does not publish benchmark claims without a reproducible method.

Newsroom →

Comments

0
  1. No comments yet — be the first.

Write a comment

Comments are public. We do not publish abuse, spam or advertising.