Earliest located milestone
Sept. 112025 · College StationHundreds filled City Hall, 5,047 petition signatures were delivered and the council unanimously rejected the proposed land sale.
Texas public decisions · 2007–2026
How Texas Built the Data Center Boom
Texas did not simply get a data-center boom. Public decisions helped build the runway—first with tax policy, then local deals and utility connections, and finally a new rulebook for the demand that followed.
Published Aug. 3, 2026 · sources inspectable
54Dated public actions and project milestones
15Projects in the jurisdiction explorer
7Local deal chains opened for inspection
Scroll to follow the chain ↓
The proposition
The boom has a public record.
Start with the decisions, not the skyline. State lawmakers created a data-center-specific tax advantage. Cities and counties added land actions and property-tax deals. ERCOT and utilities connected a new class of enormous loads. Companies then announced increasingly large campuses.
The sequence ends not with a verdict, but with governments confronting tax exposure, grid planning, groundwater authority and the limits of local control.
Evidence guardrail: this chronology establishes sequence. It does not prove that any single incentive caused a company to choose Texas, or that every proposed megawatt will be built.
01State incentivesLaws, fiscal notes and executive actions
02Grid and utilityPUC, ERCOT and connection rules
03Local approvalsZones, abatements, grants and votes
04ProjectsAnnouncements, construction and expansion
05ConsequencesTax, electric, water and infrastructure pressure
The reversal, at a glance
The shift did not happen all at once.
Local resistance entered this review’s record in September 2025. State promotion continued after that break, a contrary statewide voice appeared in January, and Abbott’s posture moved through directives, a proposed rural prohibition, a project withdrawal and—on the research cutoff date—a comprehensive audit gate.
Documented footprint
1 → 12Distinct jurisdictions · 12× cumulative countThe count reaches 12 by July 23, 2026. It measures retrieved siting conflicts, not all Texas opposition or the intensity of public opinion.
Latest Abbott action
Aug. 32026 · Audit before progressionPUC and ERCOT were directed to audit every queued data-center project and deny grid connection to projects that fail applicable requirements.
Two synchronized records · September 2025–August 2026Numbered markers are state-posture events; the stepped line is the cumulative jurisdiction count.
PromotionCountercurrentAbbott actionLatest action
Active record
2007
01 / 54
Local approval
Local approval
Pins mark jurisdictions, not parcels · city/county centroids unless noted
What happened
Who acted
Authority
Vote
Public commitment
Anatomy of a deal
The company name is rarely the whole chain.
Major projects can sit behind subsidiaries, separate zones, tax agreements, utility actions and later amendments. Open a case file to see the public sequence—and the missing records.
The resulting pressure
Four large numbers. Four different meanings.
Planning forecasts, interconnection requests, record peak demand and tax expenditures answer different questions. Collapsing them into one “impact” number would mislead.
Projected data-center load in ERCOT’s April 2025 transmission-service-provider forecast, up from 29,614 MW in the preceding forecast.
Large-load requests in ERCOT’s interconnection process. Requests are neither completed projects nor a demand forecast.
Requests considered for the new study process; ERCOT said nearly 89% were data centers.
The Comptroller’s estimated annual value of the limited sales- and use-tax exclusion for property used in data centers.
The bars compare scale only. The 85,508 MW system record is observed demand; 77,965 MW is one category in a planning forecast; 232,000+ MW and 438,000+ MW are request inventories measured at different dates.
Legislative accountability
The governing votes.
Certified chamber totals show how the state’s policy changed. This edition does not label individual current officeholders because the available evidence does not consistently resolve same-name collisions, chamber changes and special-election membership across the full record.
| Measure | What it did | House | Senate | Final status | Record |
|---|---|---|---|---|---|
| HB 1223 (2013) | Created the standard qualified-data-center sales-tax exemption. | 137–3 2 present, not voting | 23–8 | Signed June 14, 2013 | Enrolled text and votes |
| HB 2712 (2015) | Added the 20-year large-project exemption and included local sales taxes. | 114–26 3 present, not voting | 23–8 | Signed June 10, 2015 | Enrolled text and votes |
| SB 6 (2025) | Created large-load interconnection, financial-security and curtailment requirements. | 103–25 2 present, not voting | 31–0 | Signed and effective June 20, 2025 | Enrolled text and votes |
Not included in this edition: an individual “how current officeholders voted” layer. The package publishes certified chamber totals and enrolled records without making unresolved identity assignments.
Search the record
The event and incentive ledger.
Filter the full chronology by government level, company, place or policy tool. Each row preserves the vote, amount type and source instead of reducing the record to a score.
| Date | Lane | Action | Actor / authority | Vote | Amount or scale | Source |
|---|
Project and jurisdiction explorer
Fifteen places in the buildout.
Locations are city or county centroids. Status, investment, capacity and jobs remain separately labeled so an announcement is not mistaken for a completed facility.
Evidence rules
How to read what is known.
Research cutoff: August 3, 2026. Texas only.
- Sequence is not causation. Timing alone does not prove an incentive caused a project.
- Unknown is not zero. Unretrieved terms remain “not specified,” even when that makes comparison harder.
- Measure the denominator. Annual tax value, lifetime ceiling, planned capacity, request queue and system demand stay separate.
- Use the right source for the claim. Government records govern laws and public deals; company sources govern company announcements.
- Make map precision visible. Pins identify jurisdictions, not project parcels.
- Count conflicts, not sentiment. The 1-to-12 line counts one retrieved local siting dispute per jurisdiction; it is not a statewide opinion poll or a claim that opposition began nowhere else.
Research correction
HB 2712—not SB 1104—enacted Tax Code §151.3595 in 2015. The source report’s SB 1104 attribution was corrected against Texas Legislature Online before this visualization was built.