Posted by cdrnsf 1 day ago
There is no free gas pipeline capacity to get gas out of West Texas. Any time new pipelines are built, they are filled within months.
This makes a ton of sense for oil producers (which are also gas producers) who can sell their gas for less of a loss (potentially a profit!) and also for MSFT who can lock in long term contracts for minimal cost. I'd guess these contracts are for $1-2/MCF which is win/win for the oil companies in the area and MSFT.
Further, there's a gas turbine shortage so Microsoft choosing to put their (presumably limited) allocation of gas turbines in West Texas, where they have good alternatives, seems a bit mysterious. Why not save that massive amount of turbines for the northeast DCs, where renewables work far poorer yet gas is more reliable?
The reasons that seem most convincing to me:
1. Political environment is hostile to renewables and Microsoft doesn't want to paint a target on their back by choosing solar plus batteries, the choice others are making in West Texas.
2. Grid connection drastically changes economics, but pipelines for gas are cheap or something, so the massive cost and delay from grid interconnection simply isn't worth it
3. There are particular political favors going on with Chevron, e.g. Chevron wants gas in the area and is willing to increase MS's turbine allocation if they do it in west Texas, or Chevron is helping get around pesky local political approvals for data centers, or something like that.
The cost of gas does not seem like a justification for this, though.
With enough batteries they might get through the night, but seasonal shortages are much harder to handle that way.
I bet a carbon tax on data centers would be popular if the Democrats get back in.
Put another way if you want to store power say 1GWh for the worst month, well a solar panel provides a lot of power over even the worst month. 20MW of solar panel averaging even 40MWh/day * 30 days = 1.2GWh and cost way less than 1GWh of batteries.
Near the arctic circle Wind fill the same niche.
That can help on overcast days, or the middle of winter when days are shorter, but you still need battery capacity to run for stretches when there is no sun producing energy.
For residential its often recommended to have 3 days of storage capacity + the option for a backup generator. I'm not sure what the recommendation is for industrial projects, but I expect they would still need a backup generator. If they already have to put in a generator to handle the full load of the server farm, and they have cheap local gas, why bother with the solar? At a minimum you have both, but the generator (or power from a contracted gap generator) is a given.
A common misconception you don’t need to hit 100%, you need enough energy to make it either from that day or from prior days.
At grid scale daily production is never 0 and it’s never 100%. You can guarantee a surplus all 365 days a year, it’s just a cost vs benefit function, that’s however different from always fully charging your energy storage. If hypothetically the minimum was 99.9% of nighttime needs the odds you cover that gap the next day is extremely high to the point where a little extra storage makes sense vs aiming for 100% every day.
So now you’re just trying to optimize something for minimum cost. Utilities do this all the time with traditional generation you have random equipment failures and shifting seasonal demand. Thus they optimize maintenance schedules around seasonal demand etc.
Regardless, that buffer doesn't make a big difference in the topic here. My panel regularly show 0Wh on rainy days, PR effectively 0 as they may be getting only a few percent of real capacity.
How do you propose industrial scale would allow a series of arrays could both not be oversized on sunny days and cover usage on cludy days?
If I'm misunderstanding your argument I apologize. I just don't see how a data center could possibly be okay with having only a single day worth of storage and generation without having backup power scaled to cover their full usage demands.
What do you mean? Long periods of calm weather during cold temperatures is not unheard of in Finland and does cause issues at times due to amount of wind energy that has been built in recent times.
Beyond that wind for each location is somewhat seasonal so if you’re doing this for winter energy you pick areas with better wind in winter.
If you’re proposing seasonal storage as a viable option then cutting total storage needs by 90% (to handle a 10 day lull vs 90 day gap) is inherently cheaper.
Hell traditional hydro is well suited for such situations as annual output and storage is limited but nontrivial. Doubling output for 10 days isn’t nearly as problematic as trying to cover a 90 day shortfall.
Nothing in my post suggests 1GWh of storage is cheap.
Given the state of the power grid in Texas, this could be the most important consideration. Why? Texas is not connected to the national power grid, and only electricity from plants operating in Texas is available. The last winter and summer, demands on the grid have severely stressed, as reported in many places. In 2021 there as a state-wide crisis and almost a complete failure. https://en.wikipedia.org/wiki/2021_Texas_power_crisis
In short, you're going to have to build your own power plant anyway, so why bother with the grid? Gas is the cheapest, fastest zero-to-production choice for onsite power generation, and has been for a long time. Unless you're dealing with nuclear, the fuel cost just doesn't matter compared to the rest of the buildout, and gas wins because you can take off-the-shelf turbines and bolt them down.
You can only get away with it in places that don't care about environmental regulations, which are the places most likely to approve new buildout of gas infrastructure anyway. Nobody in the northeast is going to approve the creation of a brand new carcinogen factory.
Is it, though? There's a ton of projects out there waiting to get their solar on the grid in west Texas, partnering with one of them and launching the project early while waiting on the interconnection queue, and adding enough batteries, gives a more robust solution right now without the SPOF nature of large single generators.
Add to that the long backorder list for gas turbines right now, with no end in sight, and I'm surprised that Microsoft would power this particular location with turbines because it's probably their best chance to do off-grid massive solar projects.
Massive off-grid solar is what China is choosing for some absolutely massive new industrial projects. Nuclear is a no-go because it takes so long to deploy, but solar + batteries are cheap COTS and available in abundance, unlike gas turbines.
gas turbines run at night too so there's no storage/backup supply issue to consider. They also take up significantly less space than wind or solar and those data centers are already gigantic.
If you want power quickly, wind, solar, and battery can be planned, installed, and producing power within a couple of years. Nothing else comes close to being that quick. If you want power this decade, that's pretty much the only thing available.
Yes.
There is no point in waiting for interconnection when you can just... not do that, and do all your generation behind the meter, with complete control of the generation to match your load. Solar wants an interconnect so they can sell off surplus; with gas you just turn the dial down to meet the load and walk away.
The clock is running on the datacenter goldrush. 70-90% of the capex window is going to be soaked up just with construction time. Introducing a capricious ERCOT permit process and shopping around for friendly solar projects to hop in bed with makes no sense when you can just write a check and solve the problem forever.
I'd bet the deal with Chevron was to enable Microsoft to hop the queue here and get those GEV turbines soonest.
Data centers would ideally run at ~100% utilization, so any drop in solar output needs to be fully met by batteries.
I don't think this is broadly accepted among major data center operators.
Now it might be that a new data center will find it more difficult to obtain a reliable supply of electricity than to obtain GPUs, but your statement was about expenses.
When you can colocate across the street from someone who's otherwise paying to get rid of it is.
When the fuel is free-ish who cares about backorder on turbines. You can run big diesels on NG (just like the pipeline people do).
Respectfully, the laws of NPV and IRR hurdles don't matter in Chinese infrastructure.
Whatever you choose in the US, it's not cheap, and developers crawl over each other to sign up hyperscalers. Race to the bottom.
If demand > supply then the price goes up. Doesn't mean you can't buy something though.
Last I heard the wait time for turbines was ~5 years at the moment. I'm sure MSFT has some inside baseball with Chevron but it doesn't work as a general rule of thumb.
To sell excess back to the grid when funny things happen to texans in the winter?
Oy vey, what a perverse incentive.
Cheap gas is great, but 2.67GW of new build natural gas in this market will cost $6-8 billion in fixed costs. You need wholesale pricing of ~$50-60/MWh ... OVER 25 years! ... to recover just the fixed costs.
For West Texas, prices averaged mid-$30s over the last year.
Microsoft has all of the leverage here, and Chevron wants a big announcement in an area where they don't have a lot of experience.
Texas is going to be somewhat shielded from this because of US policy on this front. But probably only for a few short years. Mostly LNG exports are currently very lucrative. But LNG production is bottlenecked on expensive infrastructure and shipping. And of course lots of importers of LNG are looking for more affordable alternatives as well because it is expensive. Doubly so since the recent Gulf conflict. A lot of planned infrastructure expansion just got cancelled.
So, there's probably a bit of gas overproduction happening in Texas currently and that's going to cause predictable issues when demand is going to be structurally lower. And the double whammy of petrol/diesel also going into structural decline is going to leave Texas with a lot of over production.
This won't apply to every datacenter, but the AI inference ones especially, should be seeing most demand during the day. So what's built in north America is used when it's daylight there?
If so, isn't that a perfect case for solar?
To be clear, I'm not saying it can power down, but at night it should be able to scale down significantly?
I don't know if this is necessarily true - latency isn't really important for inference in the same way as many other services (at least the max ~300ms latency you get from hitting something on the other side of the globe) - compute in NA can serve all the other timezones just fine.
Demand for AI is global (except for Anthropic, haha).
When you build a DC that works only when the sun is shining, you are wasting half of you GPU capacity
Nearly all new additions to the grid are solar, wind, and storage right now on Texas' grid. Not because of Texas regulations, but because Texas' grid is one of the few grids where generation decisions are all made by independent investors trying to make money.
Especially with the shortage in gas turbine manufacturing, very surprising! Not sure if this says more about Microsoft or datacenters.
There are cases where the fields can produce more than the pipelines can carry away. If you put your gigantic gas turbines right next to the fields you can obtain access to some extremely cheap fuel. They might even pay you to burn it sometimes. Negative gas prices are a thing.
Look at the map for 2026 of the grid buildout in Texas at the bottom of this page:
https://www.eia.gov/todayinenergy/detail.php?id=67205
All solar and batteries (yellow and black), with a few tiny blue dots for gas. It was the same story in 2025. And it will be the same story in 2027 because solar and batteries are getting even cheaper.
These are all decisions from private investors, trying to make money, and choosing solar and batteries over gas in the market where gas is the cheapest in the world, gas is like a waste product that's hard to get rid of.
Why would Microsoft choose dirty energy when all the profit-driven investors are choosing cheaper solar and storage?
It's something like 5-10 orders of magnitude cheaper to move information over fiber than it is to move the energy required to produce that same information through a [pipe/power]line.
Yes, this is weird and no, I have no idea why we do it, but it’s really weird to read “export it to east Texas” — to the extent that I had to re-parse the sentence to figure out what you meant.
Who cares if it's cheaper. It's that you're moving less of it. The more processing you can do near the source the smaller and cheaper your pipe out to the consumer can be.
Cut the tree on the hillside. Mill it in the valley. Then spend your precious boxcar volume shipping only the finished lumber out of the valley.
It’s easy to see.
Of course it’s idiotic to actively hobble clean energy. Or to put your finger on the scale for one source of energy, like the current administration does.
But it’s not crazy to argue for “energy abundance” where the market just picks the cheapest energy on the market in the US and that just gradually moves cleaner over time.
So Texas is not a laggard when it comes to clean energy, they are actually driving clean energy forward the most, because clean energy is the cheapest and most profitable energy. And that's despite Texas having natural gas that's insanely cheap right from Henry Hub.
What this tells me is that like most hyperscalers, Microsoft is not price sensitive on the electricity side, because energy costs are tiny compared to the massive capital costs of the GPUs. But why would they go this direction? What political influence would make Microsoft choose more expensive electricity, when in the past they've been fairly good at driving clean energy forward in their data center power choices, and they'd pay a premium on energy costs to go with clean energy?
edit: for example that EIA list of new solar projects you linked indicates that the largest battery installations going up in '26 are all ~500MW, and that there are only four of them (of that size). I think the energy intensity of a multi-GW datacenter is the main reason that they're not going for solar here.
sibling comment has it, they want to do power generation on site and not connected to the grid and all the PITA that come with that. Further, they can pitch power independence to the locals which removes a big argument from the anti-datacenter crowd. Finally, the power gen i saw at Stargate in Abilene TX which was maybe 10 units (if that's what they're called) took up maybe 30 acres of land so they're not very big compared to the rest of the campus.
What this should likely tell you, is that you are missing information and have an incomplete picture of the situation.
Or it could be MSFT just likes to spend extra money for no reason because they are simply stupid. I'm gonna go with the former though.
I'd be interested in all these behind the meter setups for large 24x7 loads that are being built using solar+battery though. I haven't heard of one personally, but I must be lacking information on the subject since you seem so certain these are common?
https://www.datacenter-forum.com/edora/eurowind-energy-and-e...
There are many datacenters in Denmark with decade-plus power purchase agreements from a specific wind farm, but with so much news recently I can't find if there's any operational with their own supply.
My frustration is that reliable power is the expensive part. Anyone can stand up some nameplate capacity in renewables, net out their numbers and pretend they are 100% green power in a press release. All while drawing from the "free" battery the grid supplies them at night or seasonally and letting someone else deal with the dirty part of it all. It's just not very interesting to me, as it's mostly marketing.
This one does seem to be grid-tied, but is much closer to the vision of actually being powered down to that last 5% via on-site batteries and wind. Power is like many things - the last 5% of reliability (or 1%, whatever) is the hard and expensive solve.
I remain skeptical batteries are going to end up being a solution for major datacenter buildouts in most locations. But I've been wrong before, and would love to be wrong about this one perhaps the most!
Except that there's externalities not priced in. The consequences of those are becoming increasingly visible and very expensive to address
> A majority of the generation will come from large GE Vernova (NYSE: GEV) turbines.
Perhaps Microsoft had better ability to overturn local opposition to data centers if they had Chevron's political influence over the politicians too?
Chevron and the US Government are joined at the hip, so these kind of deals "flow" naturally.
[1] https://www.microsoft.com/en-us/corporate-responsibility/sus...
Powering the future through innovative, sustainable energy solutions.
Solar Turbines Incorporated, headquartered in San Diego, California, is a wholly owned subsidiary of Caterpillar Inc. Solar manufactures the world’s most widely used family of mid-sized industrial gas turbines, ranging from 1 to 39 megawatts. More than 17,000 Solar units are installed in more than 100 countries with more than 3 billion operating hours. Solar is a leading provider of energy solutions, featuring an extensive line of gas turbine-powered compressor sets, mechanical drive packages, and generator sets.
In such a world, you would have looked it up before posting, and your comment, their reply to you, yours to them, and mine to you, wouldn’t exist.
Perhaps they were trying to send a reward signal to all of us in this thread in order to induce a shift from where we find ourselves now in order to arrive upon the “look it up before posting” world line?
When they say "large GE Verona", they mean the 7HA. This is an actual power plant with proper emissions controls. Not the aeroderivatives in parking lots we've seen so far.
> Their plan includes the use of seven U.S.-made GE Vernova Inc. GEV 7HA natural gas turbines to deliver the plant's initial capacity.
https://finance.yahoo.com/sectors/energy/articles/chevron-mi...
This is where government regulation, ideally a smart representative of the people's interests, comes into play. Hence the effort for companies to do regulatory capture. (See point #1)
- A Microsoft employee 2080
They'd also be accelerating the collapse of AMOC... after which they won't need ACs anymore.
So I guess, yeah, this solves the problem of too much heat, in a fashion!
There are also a lot of smaller nuclear power plants planned: https://en.wikipedia.org/wiki/Nuclear_power_in_Poland
We must stop using fossil fuels as fast as possible, globally.
They aren’t building data centres just for Ai but building that can be repurposed.
They can use it for research or as new location as part of Azure and so on?