Texas Halts Data Center Grid Approvals as Power Constraints Reshape Interconnect Strategy

Grid Bottlenecks Are Now the Primary Interconnect Constraint

Texas Governor Greg Abbott ordered a pause on approvals of new data center projects through the state's grid interconnection process, citing reliability concerns as electricity demand surges. For enterprise buyers with Texas expansion plans, this creates immediate schedule risk and raises the value of pre-energized sites, off-grid power arrangements, and alternative markets with clearer interconnection paths.

The pause shifts competitive advantage to operators with ready capacity outside ERCOT-constrained markets. Buyers banking on rapid utility-connected capacity in Texas now face delays, making "build-it-fast" Texas campuses less attractive relative to regions with transparent interconnection processes and available grid capacity.

New York followed with an executive order establishing a temporary moratorium on data centers while the Department of Public Service forms a Data Center Interconnection Working Group within 60 days. The working group will apply "beneficiary pays" principles to network upgrade and resource adequacy costs, making interconnection economics more explicit in procurement decisions. This increases upfront costs and approval timelines for large loads in New York, strengthening the case for colocating in neighboring states or choosing providers that absorb upgrade complexity rather than passing it through immediately.

Subsea Capacity Surge Reflects AI and Cloud Traffic Growth

More than $12 billion in new submarine cable projects for trans-Atlantic and trans-Pacific routes was reported this week, with Google, Meta, and Microsoft collectively driving the spend to support AI workload distribution across continents. The investment should improve international path diversity, but the near-term effect is longer lead times before usable capacity arrives.

Enterprise buyers with global cloud, backup, or latency-sensitive workloads should expect procurement conversations to increasingly include route diversity, cable landing risk, and multi-region resilience. This strengthens the position of hyperscalers with direct cable ownership or anchor tenancy, while raising pressure on traditional carriers and bandwidth wholesalers to prove they can match the resilience and scale of hyperscaler-built routes.

Lightstorm plans to launch an India–Malaysia–Singapore subsea cable system with Microsoft, Singtel, and Tata Communications involved. This is relevant for enterprises with Asia-Pacific footprints because it improves latency, resilience, and carrier diversity between India and Southeast Asia, two regions where enterprise cloud demand is still growing quickly. It challenges existing regional subsea and terrestrial connectivity providers by adding a new route with hyperscaler-backed demand, which improves bargaining leverage for large buyers and cloud partners.

Long-Haul and Metro Capacity Moves Signal Tighter Supply

Verizon CEO Dan Schulman said the company secured a more-than-$1 billion dark fiber agreement with Google. Dark fiber deals of this scale signal that hyperscalers are locking in route diversity and long-haul control for AI and cloud traffic. That tightens supply and makes it harder for enterprises to negotiate large-scale bespoke fiber builds on similar corridors. This favors operators with extensive owned fiber and engineering capacity, while increasing competitive pressure on wholesale providers to offer either lower-cost lit services or more flexible private-fiber arrangements.

AWS is integrating AT&T into its Interconnect service, and AT&T Business is extending 400G wavelength service to more than a dozen metros with edgeless handoff capabilities. These moves matter for enterprises buying private cloud connectivity, multicloud transport, and metro-to-long-haul bandwidth because they suggest more high-capacity options and fewer custom cross-connect designs. AT&T's 400G push puts pressure on other wavelength and DCI providers—especially Lumen, Zayo, Verizon Business, and regional fiber operators—to match metro coverage and port speeds or risk being seen as legacy options.

DriveNets unveiled 1.6T high-capacity AI fabric platforms using Broadcom silicon to connect thousands of XPUs and reduce latency for AI workloads. This matters for enterprises building AI clusters or buying interconnect for GPU-heavy environments because 1.6T-class fabrics lower the cost per bit and help postpone some internal network redesigns. It increases pressure on networking incumbents and optical vendors—especially those selling 400G/800G gear—to justify upgrade timing versus a faster move to 1.6T-class designs.

What to Watch

FERC is expected to act on large-load interconnection reform by the end of June, and PJM has been described as shifting responsibility for managing data center power demand toward states rather than the grid operator level. If large-load rules tighten, buyers may see more explicit requirements for load flexibility, curtailment, or self-supply, which changes how they budget for capacity and backup power.

Buyers should prioritize providers that can document interconnection status, grid upgrade timelines, and power availability in writing. Pre-energized sites now carry a measurable premium over greenfield builds in constrained markets. Dark fiber and subsea capacity are increasingly locked by hyperscalers, so enterprises needing global or high-capacity routes should start procurement conversations earlier and evaluate multi-carrier redundancy more seriously than in prior cycles.