Forgot your password?
Close
typodupeerror
Power United States

Texas Grid Flags Risks As Data Centers, Crypto Sites Fail Voltage Tests (reuters.com) 105

Posted by EditorDavid from the offing-the-grid dept.
Reuters reports: Several large data centers and crypto facilities planning to connect to the Texas power grid ahead of peak summer demand have failed key reliability tests, raising the risk of power outages just as electricity use hits its seasonal high, according to the state grid operator... Unlike traditional industrial customers, which tend to draw electricity steadily and predictably, data centers are engineered to cut their connection to the grid at the first sign of trouble to protect their equipment and keep services running. That makes them an unpredictable and potentially destabilizing force on grids already under pressure from rising demand. Four groups of unnamed large electricity users, including data centers, abruptly disconnected from the Texas grid during a test of how they would handle routine voltage disturbances, the Electric Reliability Council of Texas (ERCOT) said in a report dated May 21.

When large customers abruptly cut their power use, it can knock the grid off balance and trigger wider outages. ERCOT, which manages electricity for most of Texas, said it reviewed about 20 gigawatts of large customers seeking to connect to the system, including eight projects totaling roughly 3.9 gigawatts aiming to start up before July 1. It said it identified four groups of large power users that could each trigger more than 5,000 megawatts of demand tripping under certain fault conditions, based on simulations of transmission system disturbances. Those abrupt drops in demand were equivalent to the electricity consumption of a large city such as Boston.
This discussion has been archived. No new comments can be posted.

Texas Grid Flags Risks As Data Centers, Crypto Sites Fail Voltage Tests More

Texas Grid Flags Risks As Data Centers, Crypto Sites Fail Voltage Tests

Comments Filter:

  • Perhaps exceptionally large power users that can quickly drop their demand should be required to have a flywheel energy storage unit on-site to smooth their cut off and start up demand.

    When they see the need to shut off their servers, they transition their power requirements to spinning up a flywheel storage system, then gradually reduce their power requirements. When they are ready to start up their servers again, they can tap into the flywheel's energy and slowly transition back to grid power.

    Using a fly

    • Even a water tower where you run pumps to store grid power and turbines to reclaim it. And it is relatively cheap to install if max efficiency is not a requirement.

    • Saw such a device in the basement of a data center that I visited 35 years ago when I was in college.

    • Re: (Score:2, Interesting)

      by dfghjk ( 711126 )

      "Perhaps exceptionally large power users that can quickly drop their demand should be required to have a flywheel energy storage..."

      Why should the customer provide the solution? Centralizing storage at the source would be more cost effective and would be under the control of the people who depend on it.

      • Re:Flywheel storage buffer (Score:5, Informative)

        by Jumperalex ( 185007 ) on Monday June 08, 2026 @07:00AM (#66180220)

        Why should all the other customers have to pay for it?!?! Also that's not how grid load balancing works.

        It is a problem specific to the large user and that same large user should be held responsible for being a responsible part of the community.

        • Why should all the other customers have to pay for it?!?! Also that's not how grid load balancing works.

          It is a problem specific to the large user and that same large user should be held responsible for being a responsible part of the community.

          The Texas grid which is separate from the rest of the US has a size issue with balancing and peaking that ha shown up from time to time. Their grid collapse a few winters ago would (probably) not have happened if they were connected to the big grid like everyone else is.

          What I'm seeing here is a very good possibility that this will end up being a year round problem for Texas.

          • The Texas grid which is separate from the rest of the US has a size issue with balancing and peaking that ha shown up from time to time. Their grid collapse a few winters ago would (probably) not have happened if they were connected to the big grid like everyone else is.

            There are two other US grids, the Eastern and Western Interconnects. They are not frequency synchronized, and there are minimal AC-DC-AC interties connecting them.

            The big collapse in Texas involved about 30GW of generation going off l

            • > The big collapse in Texas involved about 30GW of generation going off line. Even if Texas were connected to the Eastern grid, it is unlikely that there would have been 30GW of spare capacity and 30GW of available transmission to draw on.

              If Texas were connected to either of the national grids, they would have been i compliance with federal standards for their infrastructure and would likely not have had the failure in the first place. The only reason Texas hasn't joined the rest of the nation is because

              • This proposed large HVDC line (Texas to Southeast via Southern Spirit ) is designed specifically to preserve ERCOT independence. FERC approved it without triggering full federal jurisdiction over ERCOT operations because HVDC keeps the grids asynchronous (controlled, scheduled flows via converters). It's not "leeching"; it's bilateral trade that benefits reliability and markets. Texas exports power often (cheap renewables/gas); imports during peaks. Other regions do the same and none of it's free or some ki

                • > It's not "leeching"; it's bilateral trade that benefits reliability and markets.

                  So what would you call it if you say, joined in a carpool group where you took turns driving with your own vehicles, but one member had a notoriously unreliable car that you had to take extra turns in your vehicle to cover for?

                  This isn't about normal ebb and flow of energy. Texas gains a lot more out of this deal than anyone else, because their grid has lots of problems with their poorly maintained generation capacity. So r

                  • Oh, you wanted a citation. Okay, you get AI slop since you're too lazy to search your self The paper "Cascading risks: Understanding the 2021 winter blackout in Texas" (Joshua W. Busby et al., Energy Research & Social Science, 2021): "Imported power would not have fully compensated for the loss of production inside Texas, but it's possible that a few GW of additional capacity would have reduced the scope and duration of blackouts." This directly supports the point: the generation shortfalls of 34 GW

                    • Thanks for demonstrating my point.

                      ERCOT suffered nearly double the shortfall of their neighbors. Now, if you continue to read that paper there's another key point:

                      "On the supply side, Texas has undertaken some efforts to protect its grid against more common summer weather demand spikes, but it has done little to protect itself against rarer winter freezes. However, these happen often enough that more winterization is warranted. Texas was advised to winterize by FERC after the 2011 freeze led to similar, alb

                    • And absolutely zero of all that cherry picking and lame quoting about unrelated topics shows that neighboring states could have made up the shortfall or prevented the outages. It's a weak attempt to move the goalpost anywhere but where you started: no, neighbors were not going to help in Winter 2021 because they were load shedding themselves and having their own problems with the storm. You also don't acknowledge that some of the measures Texas has taken have been effective and no more outages of that scal

                    • > where you started: no, neighbors were not going to help in Winter 2021

                      Please quote where I said such an interconnect would have had any impact on the 2021 disaster. What did I say to make you think that's my position.

                      What I said was, to quote myself, "leech off their neighbors when shit inevitably goes sideways again." You do realize that rolling blackouts are like an annual thing there, right? Despite being a "net exporter" they still can't always meet demand? That the *entire premise* of this /. arti

                    • I guess you are a dementia patient who cannot remember what you said. Here you are, let me help you out grandpa: "If Texas were connected to either of the national grids, they would have been i compliance with federal standards for their infrastructure and would likely not have had the failure in the first place." I've highlighted the part where you said exactly what you just denied saying. Take your meds and remember memory care therapy is at noon. Try not to fall asleep before then, hmmmkay?

                    • Let me help you out with that.

                      "If Texas were connected to either of the national grids, they would have been [in] compliance with federal standards for their infrastructure and would likely not have had the failure in the first place."

                      There. Let me break it down for you a bit more;

                      If { connected to the national grids } then { federal compliance = true }

                      If { federal compliance == true } then { reduce_disaster_probability() }

                      I then go on to explain that the entire reason they are using the HVDC interlink is t

                    • Do you need some tapdancing shoes to go with that, maybe a tu-tu and a little organ to grind?

              • If Texas were connected to either of the national grids, they would have been in compliance with federal standards for their infrastructure and would likely not have had the failure in the first place.

                Exactly, the issue was that 30 GW loss was on non-winterized infrastructure, mostly natgas. If the infrastructure had been maintained, there wouldn't have been an issue, just normal gridswitching in an out.

                That's the part that is hard to grok. In teh north, the weather that brought Texas to its knees is just normal winter weather here. Probably because we comply with regulations and take decent care of our infrastructure. And true to form, we have some guy in here claiming I am a left winger and I h8t T

                • It wasn't solely a winterization issue. Had the entire state been running on natgas, the outages would have been near to nonexistent. It was the poor winterization combined with the massive drop in pressure along the natgas lines dropping the internal temperature of the lines precipitously as every available turbine spooled up to try and compensate for wind and solar power shitting the bed during the event.

            • The big collapse in Texas involved about 30GW of generation going off line. Even if Texas were connected to the Eastern grid, it is unlikely that there would have been 30GW of spare capacity and 30GW of available transmission to draw on. So looks like a well planned system - 30 GW going offline? Explain how this is something that would happen elsewhere. I mean, you're defending their grid. So I want to hear how that was something that was random.

          • Their grid collapse a few winters ago would (probably) not have happened if they were connected to the big grid like everyone else is.

            Citation needed. Nobody actually thinks that but it seems you're making it up to make your anti-Texas hopes seem more realistic.

            There isn't that much interconnection because Texas (mostly ERCOT) runs its own grid to avoid federal regulation (FERC oversight) since lines don't cross state borders much. This dates to the 1930s. It gives more local control, faster renewable buildout (Texas leads in wind, big in solar), and a competitive market.

            What I'm seeing here is a very good possibility that this will end up being a year round problem for Texas.

            What I'm seeing is someone who is always lefty-style-hateful and

            • Their grid collapse a few winters ago would (probably) not have happened if they were connected to the big grid like everyone else is.

              Citation needed. Nobody actually thinks that but it seems you're making it up to make your anti-Texas hopes seem more realistic.

              You can't look up things yourself? Perhaps not. Anyhow, after the grid first went down, Your Governor blamed it on Frozen Wind turbines. Is that your. position? Regardless, your State wanted to avoid Federal oversight, and your state did not winterize the Natural gas infrastructure, Or the Turbines, though they were a small part of the problem. The drop-off in gas power due to lack of winterization accounted for 5X the drop for the turbines. So the separation form teh other grids made it difficult to imp

              • Ah, I see. You're trying to change the subject and building a straw man. 2021 Winter was mostly about Winterization failures. It's been addressed and improvements are obvious (read the the article from Texas Tribune [texastribune.org] I've already linked and you tried to ignore). You said "Their grid collapse a few winters ago would (probably) not have happened if they were connected to the big grid like everyone else is." Again, rather than change the subject as you hand-wave, I'll simply say on topic I think you are complet

                • Ah, I see. You're trying to change the subject and building a straw man.

                  You asked for the citation - I gave you an abbreviated version, and a factual link.

                  You choose to claim that is a strawman.

                  You don't know what words mean do you?

                  • You cited Wikipedia and it didn't, at all, support your claim. It didn't address the fact that other states were also load shedding: a fact of Winter 21' you keep conveniently ignoring. Wikipedia's only weak claim around power interconnections was "Texas's power grid has long been separate from the two major national grids to avoid federal oversight, though it is still connected to the other national grids and Mexico's; the limited number of ties made it difficult for the state to import electricity from ot

      • The GP post is pretty far off the mark in therms of reality, but the issue is switching a >100MW load off and later on instantly. That has a very real impact on the grid. Smart system design can let a data center ramp up and down grid load vs on-site storage/generation, but not everything uses this approach as it is more expensive.

        I did work for an industrial facility many years back that needed to ramp their 35MW primary load-- the idea that they could fully unload it and restore it within a few minutes

        • Re: (Score:2)

          by Mal-2 ( 675116 )

          While a sudden simultaneous shutdown is sometimes unavoidable, there is no reason why everything has to come back up at exactly the same time. Certainly, supercomputers of decades past didn't expect that all systems would spring to life simultaneously.

    • Re: (Score:2)

      by AmiMoJo ( 196126 )

      Batteries are probably more practical for this, and can serve as a backup for them too.

      Flywheels are used, but a better place to put them is old fossil fuel plants. They already have a lot of the infrastructure. Ireland has converted some, and I'm sure other places too.

      • The reason I thought of flywheels was because I was thinking of something that could take over the more than 5,000 megawatts of load, then slowly decrease it. Batteries would normally be fully charged and unable to absorb the gigawatt load of the servers shutting down. Chemical batteries, pumped hydro, or thermal batteries might help during the re-startup but be unable to absorb the massive load of shutting down all the servers quickly.

  • Wait, what? (Score:4, Interesting)

    by davidwr ( 791652 ) on Monday June 08, 2026 @01:01AM (#66179958) Homepage Journal

    data centers are engineered to cut their connection to the grid at the first sign of trouble to protect their equipment and keep services running.

    You are building a billion-dollar data center and you aren't putting routine-but-poor-power-quality-tolerant power-conditioning uninterruptible power supplies between the grid and your sensitive equipment???

    Besides, if you are going to build a deci-megawatt-or-bigger power consuming complex, it would help the grid out if you put some grid-scale-batteries and a large amount of always-on local power generation on-site, as some data center complexes and other heavy-industry-consumers are already doing.

    • Sorry. I need decaounces of caffeinated liquid.

    • >You are building a billion-dollar data center and you aren't putting routine-but-poor-power-quality-tolerant power-conditioning uninterruptible power supplies between the grid and your sensitive equipment???

      How will the UPS help the grid if the data centers suddenly cut their connection, e.g. stop drawing power? The local backup or power generation is likely already in place - the issue is the spike they'll produce on the grid when the data center switches to that alternative local source abruptly...

      • some DC's have power the local generation controlled by the power co as part of an joint deal.

      • Re: (Score:3, Informative)

        by davidwr ( 791652 )

        How will the UPS help the grid if the data centers suddenly cut their connection, e.g. stop drawing power?

        The issue is that the disconnections happen when grid power quality drops below acceptable levels.

        A poor-power-quality-tolerant UPS doesn't have to be disconnected from the grid when the power quality dips.

        Therefore such sudden cutoffs would be much rarer.

      • It won't help the grid at all, but by golly, the precious data center will live to serve another gigabyte of AI slop. Priorities!

      • You intentionally don't keep them at 100% charge so they can absorb power when the need arises and then give it back.

      • Dual conversion and line interactive UPSs can handle a wider range of input voltage, but have higher losses compared to offline UPS designs which still offer a 12-16ms transfer time. The secondary problem is re-connection from on-site generators which can still be staggered in smaller blocks if designed well.

    • Re:Wait, what? (Score:5, Insightful)

      by fuzzyfuzzyfungus ( 1223518 ) on Monday June 08, 2026 @02:03AM (#66180006) Journal
      I suspect it's a straightforward incentives problem. If you can get away with making it the grid's problem there's not much incentive to pay for more expensive facility power setups. Presumably this is why ERCOT is testing current and prospective customers and making noise about it; and why there are at least some standards for how ill-behaved a load can be while still being allowed to hook up; with some awkward interactions between very large sites that also have the ability to shut down rapidly at relatively low cost. If you are 'mining' crypto you presumably prefer the gear to be online because it is depreciating by the minute regardless; but the risk and inconvenience of shutting it down and booting it up again isn't particularly dramatic compared to having to cold start an aluminum smelter or something.

      • ding ding ding we have a winner. Sad I had to scroll this far for this comment.

    • Re:Wait, what? (Score:5, Insightful)

      by thegarbz ( 1787294 ) on Monday June 08, 2026 @02:20AM (#66180018)

      You are building a billion-dollar data center and you aren't putting routine-but-poor-power-quality-tolerant power-conditioning uninterruptible power supplies between the grid and your sensitive equipment???

      You misunderstood the problem. They *ARE* doing this. At the first sign of a voltage disturbance they switch from the grid to their UPS / own generators and keep running because they do care about themselves. This is the problem. During a period of grid instability having a massive load disconnect makes the instability worse. It's literally one of the ways cascade failures on the grid can occur.

      • Re: (Score:2)

        by dfghjk ( 711126 )

        "You misunderstood the problem. They *ARE* doing this."
        ARE they though? Says who? Texas?

        "At the first sign of a voltage disturbance they switch from the grid to their UPS / own generators and keep running because they do care about themselves."
        That term "voltage disturbance" is doing a lot of work creating controversy. It's implying a trivial problem when there's no reason to believe that's true. Eventually actions need to be taken, where's the evidence that what's happening is unreasonable?

        "This is the

        • Re: (Score:3)

          by pjt33 ( 739471 )

          A power company cares about a lot more than a single consumer

          According to TFS the problem isn't a single consumer: it's four classes of consumer, of which data centres is one.

        • You're absolutely right! And the power company has a duty to all their OTHER customers not to let a few shitty ones ruin it for everyone, so they get to choose who is their customer based on them not putting the rest of us at risk for failures.

          Funny how that logic spins around so easy when you're not sucking the dick of data centers. And no I don't love the power companies either; but they are doing their JOB by ensuring no new, or existing, customer can ruin it for the rest of us. THAT is how they are "doi

        • ARE they though? Says who? Texas?

          I think you too are misunderstanding what is going on here. The power company doesn't give a flying fuck about the datacentre and it's servers UPSes. The reliability tests being discussed here are grid reliability tests - load shedding due to voltage upset events. The main concern is the datacentres ability to ride through without loadshedding.

          I don't believe that's "the problem", I believe "the problem" is partisan.

          I believe you are speaking about something nothing at all to do with the topic.

          While it's true "cascade failures" must be understood and mitigated

          Yeah we should do reliability testing of the design. Oh wait isn't that ... be a good b

        • Re: (Score:2)

          by XXongo ( 3986865 )

          I don't believe that's "the problem",

          What you believe or fail to believe is of little interest.

          However: it's not the problem yet. It is the problem that they are testing to see whether the scenario is going to be a problem. The system failed the test, showing that it would be a problem, if they don't mitigate it.

          Test before you implement. Good strategy.

          I believe "the problem" is partisan.

          I have no idea why you believe that something that's clearly an engineering problem is "partisan," but really, what I said before still holds: what you believe or fail to believe is of little

      • I think the OP is saying that a better quality UPS will "improve" poor input quality, while a cheaper one will just disconnect.

        However, I don't know enough about grid stability to know whether this is reasonable. My guess would be that abrupt disconnect for over-voltage, rather than just draw less current and down convert will tend to increase the grid instability, while an abrupt disconnect for under-voltage would tend to improve grid stability.

        I would bet that the engineers at the power company have alrea

        • A traditional UPS comes in two variants, UPS and SPS, U = uninterruptible and S = standby. But anything which can supply backup power is arguably a UPS and that includes solar power systems and the like. These days it is common for them to be grid tying. And it's now common for grid tied inverters to have a boost mode, where they will compensate for voltage sags by supplying synchronized power. Therefore the functionality is absolutely available, though what exactly the hardware is called may vary.

        • I'm pretty sure a data center cutoff is bad on an under voltage too.

          1) grid under supplies and voltage drops
          2a) grid requests more power from flexible (gas) plants
          2b) HUGE customers completely cut their draw
          3) more power with less demand causes huge over voltage

      • You are building a billion-dollar data center and you aren't putting routine-but-poor-power-quality-tolerant power-conditioning uninterruptible power supplies between the grid and your sensitive equipment???

        You misunderstood the problem. They *ARE* doing this. At the first sign of a voltage disturbance they switch from the grid to their UPS / own generators and keep running because they do care about themselves. This is the problem. During a period of grid instability having a massive load disconnect makes the instability worse. It's literally one of the ways cascade failures on the grid can occur.

        Yup, And it's worse for the Turbine generating plants. Going from running a heavy load to essentially instantly a lot of the load disappearing is hell on turbines. People could think of it as going uphill in their car with the pedal to the metal, then shifting into neutral suddenly. The load on the engine drops, but at full throttle, it will rev as far as it can.

        • Going from running a heavy load to essentially instantly a lot of the load disappearing is hell on turbines.

          That isn't what happens though. Turbines can't instantly spin up and instantly take load. The primary feed trips and the entire facility switches to batteries. The batteries ride the facility through while turbines spin up, this could be seconds to minutes depending on the design, and then when turbines are up loads typically switch back off batteries sequentially.

          You choose small turbines for this precisely because this is exactly the kind of duty they are designed to handle.

          The load on the engine drops, but at full throttle, it will rev as far as it can.

          No, your analogy implies there'

        • The scale of the power draw is also a problem. When I worked in a manufacturing plant, they had standby batteries for power that would power the plant for 15 minutes or so. They had diesel backup generators that were supposed to kick in within a minute that could last days before refueling. These datacenters need their power plant to be built just as backups which would never happen.

          • These datacenters need their power plant to be built just as backups which would never happen.

            That is a silly statement given that the world is full of datacentres that not only do have 100% backup power, but there's a couple that actually have datacentres running of this permanently (see the Colossus 1 debacle). Datacentres are very much designed with power systems scaled so they can meet demand for a significant time without external power. The limiting factor is nearly always fuel supply but more than enough to handle a power outage for typically 4-8h or so.

            • That is a silly statement given that the world is full of datacentres that not only do have 100% backup power,

              When the power draw of a datacenter is in the megawatt range, you do know there are no portable diesel generators that can supply that right? These megawatt datacenters should have built their own power but alas they did not.

    • When a UPS sees that utility power is out of spec, it disconnects from the grid. Now, do that at gigawatt scale.

      • Re: (Score:2)

        by dfghjk ( 711126 )

        Already happens having nothing to do with UPS's or data centers. If you sell a service, it's on you to provide it.

        • Already happens having nothing to do with UPS's or data centers. If you sell a service, it's on you to provide it.

          In most cases you are correct. I'm sure I'll be modded troll for this, but we're talking about Texas, where the power grid was largely influenced by politics. And for a lot of politicians, technical issues are almost irrelevant.

          DDG "what's wrong with Texas' power grid", there are many interesting links.

    • Perhaps we are misreading their ultimate intent with cutting themselves off. Of course UPS is a standard to maintain system operations and continued power. The disconnect feature, isn’t for that.

      It’s more a switch abused to buy power companies.

      After they force a billion or seventy in infrastructure upgrades just to handle the massive demand for AI, perhaps they will try and sismiss themselves as a paying customer, claiming many disconnects every week are “justified.” And mains pow

      • Re: (Score:2)

        by dfghjk ( 711126 )

        And maybe all of this is complete bullshit.

        A power company needs to negotiate contracts in good faith if they will require new buildout and investment. Are we to pretend that the people involved are incompetent because it suits your narrative?

    • Re: (Score:2)

      by dfghjk ( 711126 )

      "You are building a billion-dollar data center and you aren't putting routine-but-poor-power-quality-tolerant power-conditioning uninterruptible power supplies between the grid and your sensitive equipment???"

      Exactly, and "at the first sign of trouble" you deliberately force a worst case outcome? Does anyone believe this?

    • Re: (Score:2)

      by AmiMoJo ( 196126 )

      I'm sure they have adequate protection for their equipment. The problem is everyone else's. If they switch off a large amount of power consumption suddenly, the generators supplying them are instantly putting too much power into the grid, and need to ramp down. Of course many of them can't ramp down very fast, and even things like batteries can only respond as fast as they can detect the problem happening some distance away, so the voltage goes up and the frequency wonders off.

      Brown outs are bad, voltage sp

      • The risk mitigation strategy here is far more complicated than that. For a DC to completely cut grid power and switch to onsite generation would require onsite generation to have the ability to run for extended durations in case of grid collapse, including fuel for however many days/weeks that might be. If a DC is going to have that much generation onsite they would be more likely to use onsite power as the primary, grid as secondary which is also potentially problematic for the grid. The more rational r

    • Re: (Score:2)

      by gweihir ( 88907 )

      You are mistaken. Nobody puts UPS on a data-center. Far, far too expensive. What you do it you have two grid connections as independent as possible and you are able to switch over in real-time. Alternatively or in addition, you can put in a feed for battery-storage (sometimes bolstered by generators) intended to allow clean shut-down within 15 minutes or so. The number one killer of computer PSUs in data-centers is no-warning shutdown at higher electrical load.

  • It almost looks intentional (Score:5, Insightful)

    by Jeremi ( 14640 ) on Monday June 08, 2026 @01:05AM (#66179962) Homepage

    If I was deliberately trying to cause a nation-wide backlash against data centers, I'm not sure what I'd be doing differently from what the AI companies are currently doing.

    Has nobody told them that people don't like having their lives disrupted, particularly when they don't see any compensating benefit, or even a convincing reason for having any of it? If they were to ease off the gas pedal just a bit, they could probably do a boil-the-frog and get a larger number of smaller/less-obtrusive data centers built over a longer time period, and without the voter revolts and strict legislation that are likely to hobble them now.

    • I recently bought a pair of wooden shoes from a store in Holland Michigan. I wonder if this is a sign

    • Re: (Score:2)

      by MTEK ( 2826397 )

      If they were to ease off the gas pedal just a bit, they could probably do a boil-the-frog and get a larger number of smaller/less-obtrusive data centers built over a longer time period, and without the voter revolts and strict legislation that are likely to hobble them now.

      The fire hose of cash is on full blast. Spend it as fast as you can-- no time for clever scheming.

    • Re: (Score:3)

      by AmiMoJo ( 196126 )

      It's short sighted too. If they installed a lot of batteries and renewables at the site, at least they would have something of value when the AI bubble bursts and the rest of the site gets scrapped.

    • The Texas power grid is a joke. https://www.forbes.com/sites/e... [forbes.com]

    • Has nobody told them that people don't like having their lives disrupted, particularly when they don't see any compensating benefit, or even a convincing reason for having any of it?

      LOL, as if 'the people' matter. Only money matters bro. The people will be inconvenienced, or die. Money doesn't care.

  • TX: We have more data-centers than CommieFornia, neener neener! You're a bunch of `ZZZzzt *poof*

  • The biggest complains from huge data centers are noise, traffic, water use, and power use/impact on the grid, without the economic benefits that a large-energy-using factory would typically bring.

    Depending on location, a bunch of smaller complexes spread over hundreds of square miles vs. one big one might have tolerable noise and traffic levels, particularly if they are in non-residential areas. If you can get the data center down to under, a few thousand square feet, you can literally disguise it as a hou [click2houston.com]

    • Except they are using natural gas for the on site generation.

    • Good points.

      nVidia is talking about paying homeowners to install a 10 GPU unit in their backyard along these lines, going highly-distributed.

      The trick with the "data center jobs" is the estimate that 70% of them will be new H1B workers so even those claims to the locals whose politicians already waived taxes is that they're looking at maybe a few dozen local jobs for a huge data center.

      It's worth watching the Tucker/O'Leary interview to see the mindset of these people. "Corruption and screwing the locals is

      • Re: (Score:2)

        by 0123456 ( 636235 )

        > nVidia is talking about paying homeowners to install a 10 GPU unit in their backyard along these lines, going highly-distributed.

        That might work in a high-trust society, but in the modern low-trust West it just means a glut of black-market GPUs as people steal the boxes and sell the contents.

        • Re: (Score:1)

          by davidwr ( 791652 )

          There will no doubt be anti-theft/anti-tampering measures in place.

          My hope is that if someone opens the case without the right key a stink bomb explodes on them.

          More realistically, there will be technical measures in place to make the computer hardware useless to thieves. Once word gets around there won't be any incentive to steal them.

  • Converting to California units, 3.9 GW, is around 1.77 Diablo Canyons.

  • Why am I not surprised. This is nothing but a large-scale money-grab. And they are in a hurry because the hype does not have that much linger before it collapses.

  • The issue is that these large customers are rapidly cutting out, which causes the problem of instantaneous excess generation capacity. The most sensible resolution to this would be to colocate temporary sink loads near them such as Fast Frequency Response (FFR) battery banks (BESS), of which there are approximately 20 GW (ignoring energy storage capacity) of deployed BESS. This means there needs to be another approximately 20 GW of FFR BESS. This is something data centers should pay for because it's a probl

    • This is something data centers should pay for because it's a problem they've created.

      i think you've missed the point of why they are locating these in Texas.

  • Rework the demand charge structure (Score:5, Insightful)

    by hwstar ( 35834 ) on Monday June 08, 2026 @11:08AM (#66180546)

    Demand charges are charges the utility imposes on commercial and industrial customers when they attach large loads to the grid and they draw large startup currents.

    Rework the demand charges to penalize dumping loads instantaneously and the problem should be solved.

    I see no reason that data centers couldn't manage their electrical loads more gracefully.

    • ^THIS
      If it's a problem, set the appropriate requirements and charge the consumer enough for each abrupt load reduction so they'll proactively invest in solutions to keep the load reduction slope under the levels triggering penalties.

  • That's the joke I was looking for, but Slashdot failed again. Either at the comment level or the moderation level, and I don't even care enough to search to find out which.

  • Oh, yeah, a building with 1000 servers and 25 staff members, running subscription-based services will shut their whole entire operation down every time the voltage dips when it's hot. That sounds realistic to everyone who works in IT infrastructure. We can't even do that at my company and we have 10 servers!!!
  • What a nice metaphor! When the grid is in trouble, datacenters only think about their own safety and sacrifice the entire grid to keep their servers in pristine condition.
    Time to grow up and act like a real industry. This isn't solved with a new algorithm. This is way beyond ones and zeros. Time to get those keyboard hands dirty. High voltage protection gloves, anyone?
  • TCBRs are a standard grid component exactly for this purpose. The power engineers of the grids should not be connecting data centers to their grids unless they have used these as part of the circuit or required the data center to incorporate them. This is just plain engineering malpractice. They know what these data centers are drawing and they know what that means in terms of risk to the grid.

    These $500k components make the most sense as being owned by the grid because they can cover multiple data cen
  • Summer / winter is a crap for Texas power grid !

    • Re: (Score:2)

      by Targon ( 17348 )

      That is because regulations and cooperation with neighbors are against what Republicans believe in.

  • One thing about the politics in Texas, you have the politicians who are so "pro-free market", that the idea that government should regulate things and prevent corporate entities from harming the general public is against what they stand for. This is full-scale Republican for allowing the public to suffer as long as corporations are more profitable. Yea, power outages in the cold winter months and hot summer months with politicians in Texas doing NOTHING to prevent future problems after each crisis.

    Now,

Slashdot Top Deals

"Plastic gun. Ingenious. More coffee, please." -- The Phantom comics

Close