U.S. data centers now have 56 GW of behind-the-meter gas plants in development to skip 5-year interconnection queues — but GE Vernova’s 116 GW turbine backlog shows the new bottleneck just moved.
DRAM and HBM prices jumped 90-95% in a single quarter, pushing GPU lead times past 52 weeks for non-priority buyers — here is what it means for anyone planning data center capacity in 2026.
Large power transformer lead times have stretched to 3-5 years and OEM backlogs top $176B — not permits or chips, but grid equipment is now the hard constraint on data center go-live dates.
FERC ordered all six U.S. grid operators to justify or rewrite large-load interconnection rules by August 17, 2026 — a rare regulatory sprint aimed at the queue delays blocking new data center capacity, with the national queue now topping 2,060 GW.
Meta’s new petabit-class Petal cable doubles transoceanic capacity, but with Google, Meta, Microsoft and Amazon now consuming up to 90% of trans-Atlantic bandwidth, network capacity is becoming the next constraint for data center buyers.
Virginia’s new Tier 4 rules and a 2.8 GW Bloom Energy-Oracle fuel cell deal signal a real shift away from diesel-only backup power for data centers in 2026.
Data centers consume 17.4 billion gallons a year directly for cooling, and roughly 211 billion gallons indirectly through power generation. Minnesota, South Carolina and four other states just turned water access into a formal permitting gate for new sites.
Off-balance-sheet leases now total $662 billion across the top five hyperscalers, and private credit to AI firms has topped $200 billion. Here’s what that means for anyone signing a colocation or capacity contract.
Global data center insurance premiums are set to hit $24.2 billion by 2030 — but hyperscaler capex is growing even faster. Here’s what buyers need to know before they sign a lease.
45 U.S. data center projects worth $68 billion were blocked or delayed in Q2 2026 alone. Here’s what the power grid and local opposition mean for your deployment timeline.