TL;DR
X argues data centers are stealing power; ERCOT and PJM records show queues, deadlines, megawatts, and studies.
MSM Perspective
MSM emphasizes grid strain, data-center demand, and reliability planning.
X Perspective
X treats large-load queues as proof that tech companies are jumping the line.
The data-center grid story becomes less mystical when it is forced into a queue.
ERCOT's large-load integration page says entities wishing to interconnect load facilities of 75 megawatts or greater through the Batch Zero process should consult PGRR145. It lists forms, attestations, and deadlines: several interconnecting large-load entity forms are due to transmission or distribution providers by July 10, and transmission providers must submit specified materials to ERCOT by July 24. [1]
That is the opposite of the viral version. On X, a data center is an outrage object: a warehouse of GPUs stealing power from households, or a national-security asset being strangled by regulators. Mainstream stories usually land on the same grid-strain headline. The queue record turns the dispute into rows, deadlines, forms, and studies.
ERCOT's June 18 release says the Public Utility Commission of Texas approved Batch Zero for large-user connection requests, grouping qualified projects of 75 MW and greater into a single study so ERCOT can assess future demand at once, allocate available grid capacity, and identify needed transmission upgrades. The release says ERCOT is tracking more than 438,000 MW of large-load requests, nearly 89 percent from data centers. [2]
PJM's parallel problem sits on the supply side. Its April 29 Inside Lines report says 811 new generation projects, capable of 220 GW, applied to connect through the first cycle of PJM's reformed interconnection process. PJM says demand growth is driven largely by data centers, and that demand is outpacing new supply. [3]
Those records do not vindicate any easy camp. They show that Texas is batching large loads before connection, while PJM is trying to move generation through a first-ready, first-served process. One queue asks how much new demand the grid can reliably accept. The other asks how quickly new supply can clear studies and actually get built. [2][3]
The useful question is not whether data centers are good or bad. It is whether each proposed megawatt can be tied to a location, a transmission provider, a study, a financial commitment, and a delivery date. If it cannot, it is not a grid plan. It is a press release wearing a hard hat.
-- DARA OSEI, London