Texas Puts 474 GW of Data Center Demand in the Queue — and Asks for a Reveal

474 gigawatts. That is the number currently sitting in ERCOT's interconnection queue, and it is not a projection. It is more than five times the state's record peak demand, and roughly 90 percent of those requests are data centers. Governor Greg Abbott has ordered a pause on data center approvals. The Public Utility Commission of Texas and ERCOT are now auditing every facility moving through the interconnection process. Any project that fails state requirements will be denied a grid connection. This is not a market correction. It is a structural intervention. The move follows months of public backlash. In July, New York enacted the first statewide moratorium on hyperscale data centers. Around a dozen states have proposed similar bans, according to CNN. Gallup finds 71 percent of Americans oppose a data center in their local area. A Reuters/Ipsos survey puts the opposition at 57 percent. The political signal is clear. The technical question is what happens next. Abbott's five disclosure categories are public funding, power use, water consumption, community impact, and ownership. Companies must reveal taxpayer-funded incentives, projected power demand, on-site generation plans, water sources, reuse methods, and measures for noise and traffic. On the surface, this reads like a transparency measure. In practice, it is a new gatekeeping mechanism for the Texas grid. I have spent years auditing energy contracts for mining facilities, and the first thing I look for is not the number on the demand page. It is the difference between capacity and sustained load. ERCOT's queue is filled with capacity requests, not commitments. A 100 MW interconnection request does not mean a 100 MW baseload consumer. It means a developer reserved the right to consume up to 100 MW at some point in the future. That is not a load profile. That is an option contract. The disclosure mandate forces the option to be priced. Power use must align with on-site generation and real water sourcing. The crypto-mining sector has an unexpected advantage here. Proof-of-work miners are built to curtail. They can shed load in milliseconds. AI data centers need continuous, predictable power. The grid risk is not the same, and treating both as one category creates a false consensus. This is why the water disclosure matters more than it seems. A hyperscale facility with evaporative cooling can consume millions of gallons per day during a Texas summer. The grid and the water system are coupled through heat. The state is right to ask for reuse methods. But the verification question remains: who measures the cooling tower blowdown, and who checks groundwater permits? Those are not PUCT questions. The most important disclosure is the least discussed: ownership. The state wants to know who is behind each facility. That should be table stakes, but ERCOT's interconnection process was built for generation resources, not load. Now it is being asked to act like a bank doing KYC on every gigawatt-hungry tenant. That is a systems engineering problem, not a press release. I have seen this pattern before, in protocol audits. A smart contract rarely fails because the math is wrong. It fails because an assumption about the external world was encoded into the state machine without being tested. The same applies to the Texas grid. The assumption was that data centers would bring jobs and tax revenue, and that cheap Texas power would run them. The external world changed. Public funding, water, noise, traffic, and ownership all have a way of becoming real liabilities when a market turns from bull to bear. Math doesn't care about public sentiment. The arithmetic of 474 GW against a peak load below 90 GW does not resolve itself through announcements. The queue has become a speculative vehicle. Developers can submit requests without a firm offtaker, without financing, without a construction schedule. The real cost is not the energy consumed. It is the distortion of transmission planning and the artificial pressure on wholesale prices. Texas is now trying to fix that distortion with disclosure. It is a reasonable first step, but it is not a solution. Disclosures only work if the operator has the capacity to verify them. ERCOT does not. It is a grid operator, not an audit firm. The PUCT does not have the staffing to trace the ownership structures behind a dozen shell entities. These are not failures of intent. They are failures of architecture. This is the part that gets lost in the AI backlash. The 474 GW number is not proof that AI is about to consume the entire state. It is proof that the interconnection queue is no longer a serious planning instrument. It is a de facto options market with no collateral requirements. The disclosure list answers "who," "what," and "where" — but not "how long." The five disclosures are an attempt to add margin to that market. The question is whether the margin call will ever be enforced. The common phrase in Texas is simple: don't tell me, show me. In my audit work, regulation fails for one simple reason: the enforcement mechanism is separated from the data by too many layers. ERCOT knows power. The PUCT knows policy. Neither knows how to validate water-reuse claims or community-impact models. That is the blind spot. The contrarian view is that the backlash targets the wrong problem. Data centers are not the first industry to place large loads on the grid. The issue is speed and lack of skin in the game. Texas should require performance bonds for every application. A 50 MW request should carry a deposit that is forfeited if the project misses commercial operation. That would clear the queue faster than any disclosure form. The performance-bond approach has precedent. ERCOT already penalizes generators that miss commercial operation dates. Extending that to load-side requests would not require new legislation, just a tariff change. The state chose disclosure instead, which suggests the goal is political defensibility, not queue integrity. Smart contracts execute. They don't deliberate. The Texas grid is not a smart contract, but it is being forced to behave like one. The five disclosures are the rule set. The audit is the transaction. The denial of connection is the revert. But the most important variable is missing: the oracle. Who verifies the disclosures? Without a reliable oracle, the state machine will execute on flawed inputs. The AI and crypto narratives have converged at the grid connection point. Both rely on cheap energy, both are politically unpopular, and both are now subject to community governance in real time. The era of private agreements between developers and utilities is ending. Liquidity is an illusion until it is tested. Grid capacity is the same. The 474 GW queue looked like opportunity in the bull market. In a bear market, it looks like liability. Texas responded with paperwork, and that paperwork will be the template for other states. The takeaway is not that data centers are doomed. It is that the era of unchecked interconnection is over. From here on, every gigawatt will be audited before it is approved. The question no one is answering is who audits the auditor. When ERCOT has to decide between a 100 MW data center and a 200 MW renewable project in the same substation, which one gets denied first? That decision will not be made by math or by disclosure. It will be made by politics. Texas is about to learn whether its grid can survive that test.