LecturaInvestigaciónRadarMarco de inversión
Iniciar sesión / Registrarse
← Volver a la lectura
NoticiasTema · AI数据中心电网瓶颈Material2 min de lecturaEditorial de DisconfirmAI

GlobalData: Grid Delays Push Data Center Construction Timelines Beyond Two Years

GlobalData reports that the average wait time from announcement to groundbreaking for new data centers has risen to over two years, up from 18 months, primarily due to grid connection bottlenecks.

2026-11-06 18:00 GMT+8

According to a report by Data Center Knowledge citing market research firm GlobalData, grid access difficulties have emerged as the primary obstacle to expanding artificial intelligence infrastructure. For data center projects that broke ground between the third quarter of 2025 and the third quarter of 2026, the average wait time from announcement to actual construction start exceeded two years. This contrasts sharply with approximately 18 months for projects initiated between early 2020 and mid-2025.

Martina Raveni, a senior analyst at GlobalData’s Strategic Intelligence team, identified the grid as the "hardest to fix" structural bottleneck, noting that while a data center takes one to three years to build, new grid infrastructure requires five to 15 years. In major European data center hubs, developers may face waits of up to a decade for a connection. Additionally, shortages of skilled labor, such as electricians, have pushed some delivery schedules back by around a year.

Community opposition also plays a significant role. Ya-Han Brownlee-Chen, Vice President of Data Center Strategy for DP World Americas, stated that local concerns regarding electricity bills, water consumption, and noise are impacting permitting and construction schedules. This tension complicates site selection and makes construction timelines harder to predict.

While these physical constraints are clear, opinions differ on whether they have already limited AI development capabilities. Luis Fernandes, Senior Research Manager at IDC, argued that there is no hard evidence yet that compute capacity itself is a limitation, attributing rising computing costs more to material shortages, such as memory supplies for high-end GPUs. However, Fernandes acknowledged that the lack of grid connection halts entire projects, leading to stalled sites that continue to bleed cash, potentially resulting in project acquisitions or transfers in the future.

Iniciar sesión / Registrarse
LecturaInvestigaciónRadarMarco de inversión

Fuentes

  • datacenterknowledge.comDataCenterKnowledge↗
Tema

AI数据中心电网瓶颈

AI数据中心扩张的最大瓶颈已从资本或芯片供应转向电网接入能力。高盛与GlobalData等机构指出,全球数据中心电力需求预计2030年将激增170%,但美国电网平均接入延迟长达5年,韩国拟出台专用电网标准亦引发项目延期担忧。为应对“幽灵”需…

4 pasos4 piezasActualizado 2026-11-06 09:13

Cómo llegó hasta aquí

  1. 2026-10-21高盛发布报告预测,2030年全球数据中心电力需求将增长170%,美国电网接入平均需5年,成为AI扩张最大瓶颈。etnews.com ↗
  2. 2026-10-21TechRadar报道,美国公用事业公司为剔除投机性“幽灵”数据中心,要求提供财务担保和长期合同,导致约50%的规划项目被推迟或取消。techradar.com ↗
  3. 2026-10-21ZDNet Korea报道,韩国气候部拟年底出台AI数据中心专用电网接入标准,议员警告此举叠加LNG直购禁令,将使已开工的GW级项目面临设计变更与投资延期风险。zdnet.co.kr ↗

Lo que aún está pendiente

  • 美国公用事业公司收紧准入后,实际落地的数据中心容量与申请量之间的差距是否会持续扩大?
  • 韩国拟议的专用电网标准具体技术门槛何时正式公布,对现有在建项目的追溯适用性如何界定?
  • 科技巨头通过自建电源或购买碳信用抵消电力短缺的策略,能否在2027年前有效缓解局部电网压力?

También en este hilo

  • South Korea Plans Dedicated Grid Standards for AI Data Centers; Lawmaker Warns of Delays for GW-Scale Projects2026-10-21
  • US Utilities Tighten Grid Access Standards for AI Data Centers as Half of Planned Buildouts Face Delays or Cancellations2026-10-21
Leer el tema completo →
Continuar investigando →