Markets
Data loading…
Live
News
No recent dispatches

Digital Realty acquires Blackstone's data centers in Virginia for $3.5 billion: the major consolidation of AI real estate

Ongoing story : Regulation of AI Data Centers: Legislative Risk and Energy Constraints· Part 8/10

AI & EnergySubscribers only Jun 30, 2026 at 16:477Add to bookmarks

Digital Realty acquires Blackstone's data centers in Virginia for $3.5 billion: the major consolidation of AI real estate
Brett Sayles · Pexels

The most significant transaction of the year in digital real estate: Digital Realty acquires three Blackstone data centers in Virginia for $3.5 billion, setting a benchmark at $1.17 billion per site in a market with frozen permits and GPUs at 85% utilization.

Context

On June 30, 2026, Digital Realty (DLR) finalizes the acquisition of three Blackstone data centers in Northern Virginia for $3.5 billion (Data Center Dynamics, 06/30/2026). The transaction consolidates the REIT in Loudoun County's "Data Center Alley"—a corridor handling ~70% of global internet traffic. It occurs as GPU utilization reaches 85% (RUM Group, 06/30/2026), confirming the structural capacity strain already highlighted by the federal Moratorium Act (AOC/Sanders, 06/25) and Exelon's blackout alert (FT, 06/27).

Data

  • $3.5 billion: DLR-Blackstone deal value, or ~$1.17 billion/site—new market benchmark
  • 85%: GPU utilization rate (RUM Group, Seeking Alpha, 06/30/2026)
  • ~945 TWh: global data center consumption in 2026 (IEA Electricity 2025), US share ~40% (~380 TWh)
  • Digital Realty: ~300 facilities across 50 global metros
  • Northern Virginia: highest global concentration of data center capacity

Analysis

Three dynamics converge: (1) AI demand requires >100 MW facilities, which only privileged geographic corridors (energy, fiber, regulation) can accommodate; (2) Blackstone repositions itself on next-generation hyperscale data centers; (3) DLR consolidates in a market with frozen permits—several Virginia counties block new constructions due to grid constraints (Dominion Energy/PJM). At 85% GPU utilization, existing capacity is structurally strained, justifying the premiums paid for operational assets.

Probability-weighted scenarios

  • Accelerated sector consolidation (60%): Equinix and Iron Mountain follow; DLR maintains its valuation premium and captures "AI infrastructure" flows
  • Regulatory blockade (30%): The Moratorium Act extends to existing asset acquisitions—partial slowdown
  • Local energy saturation (10%): Dominion Energy delays power supply to acquired sites, deferring revenues

Portfolio implications

DLR provides direct exposure to AI capex via a liquid REIT. At $1.17 billion/site, the transaction sets an implicit valuation floor for the sector. To be integrated into an "AI infrastructure" allocation alongside semiconductor equipment manufacturers (ASML, Broadcom) and data center energy operators.

Risks & blind spots

  • Increased debt in a hawkish Fed context: the $3.5 billion debt is sensitive to real rates
  • Risk of Moratorium Act or local moratoria (precedent in Eagan, MN—court decision within 3-6 months)
  • Dominion Energy delays: acquired assets may remain underpowered

To monitor

  • DLR Q2 2026 results: capex guidance and post-acquisition occupancy rates
  • NERC July 2026 report: PJM/Dominion grid reliability
  • Congressional vote on Moratorium Act (fall 2026)
  • Eagan, MN court decision (3-6 months): AI data center moratorium precedent

Key takeaway

At 85% GPU utilization, the Northern Virginia data center market is structurally constrained, justifying premium valuations for operational assets and accelerating sector consolidation.

[/ENCADRE]

Content reserved for members

Create a free account to access all our content and the weekly review.

Article produced by artificial intelligence, reviewed under human editorial control.

Our newsroom
Was this article helpful?

9 people liked this article

Like
Arjun MehtaAnalyste infrastructure IA & énergie (Bangalore / San Francisco)
Il suit l'infrastructure de l'intelligence artificielle : calcul, data centers et contrainte énergétique.
Share:
Comments (7)

Sign in to join the discussion.

EconEddie_89 01 Jul 2026 · 04:12

1.17M per MW? Cool story-until the next utility rate hike turns these assets into stranded liabilities. Data’s not the new oil, it’s the new subprime.

le_sceptique_financier 30 Jun 2026 · 19:16

Permettez-moi de douter : 3,5 Md$ pour des hangars à serveurs, alors que même Kubrick n’aurait pas osé un tel budget pour *2001*. La Virginie finira-t-elle en cimetière high-tech comme Détroit en son temps ?

经济小王_沪 30 Jun 2026 · 13:43

35亿美金砸下去,AI基建泡沫还能撑多久?能源成本上涨和需求不确定性才是真正的考验

eco_visionario 30 Jun 2026 · 13:11

3.500M$ por 3 data centers sugiere un múltiplo de 30x EBITDA. ¿Sostenible si los costes de energía suben un 20% anual en Virginia? El riesgo no es la demanda, es el margen.

CurioBretagne 30 Jun 2026 · 12:58

La Virginie capte 70% des data centers US, mais les coûts énergétiques locaux explosent. Un pari risqué sur le long terme ?

Cla1re 30 Jun 2026 · 15:08

Et si les énergies renouvelables locales devenaient leur atout maître ?

kenji_osaka 30 Jun 2026 · 12:27

3.5B$の取引だが、AI需要のピークアウト後に過剰供給リスクはないのか。テナントの長期契約が鍵になる。

Cla1re 30 Jun 2026 · 12:15

Un pas de plus vers la concentration des infrastructures IA, mais à quel coût pour l'innovation locale ? Les fintechs africaines montrent qu'on peut faire autrement.

Topics
Explore
Information