Dallas Just Overtook Northern Virginia as the World's #1 Data Center Market. What Happens When Texas Runs Out Too?
According to Cushman & Wakefield's 2026 report, Dallas-Fort Worth has surpassed Northern Virginia to become the largest data center market in the world. The shift is symbolic — but above all industrial. It reflects exactly what GridReadiness has been documenting: when one market closes, capital moves to the next available option. The question is whether Texas is a solution or the next bottleneck in formation.
The data from ERCOT — Texas's grid operator — suggests the latter.
THE NUMBERS THAT DEFINE THE TEXAS GRID SITUATION
ERCOT all-time peak demand record: 85.5 GW
Large-load interconnection requests filed: 450 GW — more than 5× peak capacity
Projects actually approved and moving forward: ~7.5 GW
ERCOT's own CEO (Pablo Vegas): the 228 GW by 2032 estimate was "too high of a figure based on realistic expectations"
Dallas County alone: projected 10 GW of large loads by 2030
ERCOT planned transmission upgrades to relieve Dallas constraints: $14.9B through 2030
Summer 2026 peak demand forecast: 90.5–98 GW vs preliminary 112 GW figure
ERCOT's own assessment: new large-load projects "unlikely" to ramp fast enough this year
Source: ERCOT April 2026 long-term load forecast filing · Belfer Center April 2026
The mathematics are straightforward: 450 GW of requests against 85 GW of peak capacity means that more than 80% of the projects currently in the Texas interconnection queue will never receive a grid connection on the timeline their developers expect. Texas has reproduced the Northern Virginia problem at larger scale and with a grid architecture that carries additional reliability risks.
TEXAS SENATE BILL 6 — THE KILL SWITCH
Texas's legislative response to grid stress confirms the structural nature of the problem. Senate Bill 6, passed in June 2025, introduced several requirements that materially change the risk profile of Texas data center deployments:
Remote curtailment authority: ERCOT can cut power to data centers during grid stress events. Large loads connecting after December 31, 2025 must install remote disconnect equipment — the "kill switch."
On-site backup mandate: Facilities above 75 MW must install on-site backup generation capable of meeting 50% of load. This adds significant capex to every project above the threshold.
Transmission cost allocation: Data centers must pay their share of transmission costs. Current rules allowed shifting costs to other users — this is being corrected by the PUC, with detailed rules due December 2026.
Status: SB 6 is the floor. The PUC is still writing detailed implementation rules with a December 2026 deadline. The regulatory environment will tighten further.
No equivalent regulatory constraint exists for data centers connected to the French RTE grid. France operates under a stable regulatory framework where large load connections are governed by defined RTE procedures — and the fast-track process officially launched in June 2026 carries explicit government backing with no curtailment provisions.
THE OFF-GRID PIVOT — WHAT IT ACTUALLY SIGNALS
The most revealing data point in the Texas situation is the rapid growth of behind-the-meter (BTM) and off-grid deployments. Operators like Meta, Microsoft, and now Google are building data centers paired with dedicated gas turbines, co-located generation assets, or direct power purchase agreements — bypassing the ERCOT grid entirely.
This off-grid pivot is not a solution. It is a market signal that grid-connected power in Texas has become unreliable as a planning input for large AI workloads. When well-capitalised hyperscalers conclude it is faster and more financially predictable to build their own generation than to wait for a grid connection, the grid bottleneck has become structural.
Gas turbine on-site generation (Texas approach): $8–12/W installed · gas-dependent · SB 6 backup mandate still applies
Bloom Energy SOFC (Nebius US approach): $8–10/W · 90-day deployment · natural gas or hydrogen
France RTE brownfield connection: nuclear baseload at €50/MWh · 24/7 · 51 gCO2e/kWh (UNU-INWEH 2026 · rank 3rd globally)
Texas ERCOT grid carbon intensity: 350–450 gCO2e/kWh (gas + wind mix)
For operators with 2030 carbon commitments (Microsoft, Google, Meta): Texas off-grid gas is not a compliant solution regardless of cost. France nuclear is the only large-scale option that satisfies both the power constraint and the carbon constraint simultaneously.
DALLAS #1 — THE CORRECT READING
Dallas overtaking Northern Virginia does not mean Texas has solved the problem that made Virginia unworkable. It means the problem has moved states. The underlying constraint — grid connection wait times measured in years, transmission congestion, regulatory uncertainty — is present in Texas as it was in Virginia, at larger scale and with the additional reliability risks of ERCOT's energy-only market design (no capacity payment mechanism, vulnerability to extreme weather events as demonstrated by Winter Storm Uri in 2021).
The correct reading of Dallas #1 is this: capital is rational and mobile. When Virginia closed, it went to Texas. When Texas closes — and the ERCOT data shows this is already happening at the margin — it will go to the next available market with confirmed grid capacity, stable regulatory framework, and competitive power costs. That market is France.
THE FRANCE COMPARISON — UPDATED JUNE 2026
France RTE fast-track (5 official sites):
250 MW in 2 years · 1 GW in 4 years · government-backed · no curtailment risk
4,800 MW total confirmed capacity · SoftBank, Nebius, Ardian already deploying
France brownfield HTB:
18–36 months · existing HV infrastructure · 40+ proprietary sites · 10–500 MW
Texas ERCOT (Dallas):
Queue: 450 GW requests · 7.5 GW approved · curtailment risk (SB 6) · kill switch
Off-grid alternative: viable but $8–12/W + 350–450 gCO2e/kWh carbon liability
Northern Virginia / PJM:
7–10 years · effectively closed for new large loads
Ireland · Netherlands · Frankfurt:
Moratoriums until 2028–2030 · closed
The capital that built Dallas #1 is already evaluating the next move. GridReadiness tracks the grid capacity, transformer procurement windows, and brownfield site availability that determine where that capital lands. The monthly brief is the primary independent intelligence source for developers, funds, and operators making that decision before 2028.
Sources: Cushman & Wakefield 2026 Global Data Center Report · ERCOT April 2026 long-term load forecast · Belfer Center April 2026 · E&E News June 2026 · Texas SB 6 (June 2025) · RTE/Choose France June 2026.