Перейти к содержимому
Новости

US Large-Scale Data Center Pipeline Could Reach About 280 Sites by 2030

Редакция ЦОД:Главное

Источник изображения: первоисточник

The United States has approximately 90 operational data centers with power capacity above 100 MW, according to an estimate cited by TradingKey. If all currently planned projects materialize, that count could rise to about 280 by 2030, reinforcing projected demand for memory, enterprise storage and core data center infrastructure.

Главное
  • Approximately 90 US data centers are cited as operational with power capacity above 100 MW.
  • The count could rise to about 280 by 2030 only if all planned projects materialize.
  • Technology companies are expected to spend about $7 trillion on data centers by 2030, according to the cited estimate.
  • More than $1.5 trillion of data center lease agreements reported for Amazon, Microsoft, Alphabet and Meta includes commitments not yet in effect.
  • Memory and enterprise-storage demand could expand alongside power, cooling, server and network requirements.

The US market currently has approximately 90 operational data centers exceeding 100 MW of power capacity, according to The Kobeissi Letter estimate cited in a TradingKey analysis. The figure is presented as a market-wide count rather than capacity at a defined project phase or a measure of installed IT load.

TradingKey said the number could reach around 280 by 2030 if all currently planned projects are delivered. This is a conditional pipeline estimate, not an indication that those facilities are under construction, commissioned or operational.

The article also cites an expectation that technology companies could spend approximately $7 trillion on data centers by 2030. Separately, Amazon, Microsoft, Alphabet and Meta have signed more than $1.5 trillion in data center lease agreements since the start of the AI boom, including commitments that have not yet taken effect.

For infrastructure suppliers, the projected buildout could support demand across the data center stack. In addition to power infrastructure, cooling, servers and networking, larger AI deployments require substantial memory and storage resources. TradingKey identifies Micron, which participates in DRAM, HBM and NAND, and SanDisk, which focuses on NAND and enterprise storage systems, as potential beneficiaries of this demand.

The implications depend on delivery of the planned projects and conversion of contractual commitments into active deployments. Lease commitments should not be treated as operational leased capacity or as completed investment, and projected spending is not equivalent to confirmed expenditure.

Почему это важно

The projected expansion of large AI-oriented data centers could extend demand beyond compute hardware to HBM, DRAM, NAND and enterprise SSDs, alongside power delivery, cooling, servers and networking. However, the cited expansion and spending figures are forward-looking, while a portion of reported lease commitments has not yet taken effect.

Источники

Материал подготовлен редакцией ЦОД: Главное на основе указанных первоисточников.