[flagged]

Today I learned a thing! It makes sense that subsonic engines and supersonic engines would be different in retrospect but upon reading the headline I thought for sure it was going to be some kind of weird "jump on the AI hype train" article.

Good for them for trying to find a profitable proving ground for their engines.

It is at least 50% AI slop.

Siemens power-generating turbines are designed for -50C/+50C temperature envelope. All jet engines lose efficiency at higher ambient temperature due to thermodynamics, no matter how good their HP turbine blade tech.

> About three months later, we had a signed deal for 1.21 gigawatts and had started manufacturing the first turbine.

Great Scott!

1.21 jigowatts? Great Scott! the only power source capable of generating 1.21 gigawatts of electricity is a bolt of lightning

all this for predictive text, not even robotics.

Selling shovels

Great, that's what we need. More fossil fuel powered, CO² emitting, supersonic turbines polluting our environment. Unless I see a sea of solar powered carbon capturing machines,somewhere in the Saharan desert, churning the CO² back to natural gas to power these turbines, I hate this.

> I texted with Sam Altman—who confirmed power was indeed a major constraint.

Such a cheap flex right up-front, and with an em-dash to boot. I get it, it's powerful to boast about such a connection. It's just not very classy.

How loud are these turbines? Where will they be used?

Managed to talk about china's energy buildout _without_ mention of renewables? I think this pivot is 100% designed to get government money: - natrual gas turbine - china is scary - something something it's a race - china energy is good because no regulations, totally not because they are lapping the world on renewable buildout

Burning more fossil fuels in noisy, polluting ways is not a good tradeoff considering most “AI” itself is questionably a net positive, and certainly not worth the current levels of investment.

Hear me out... we could just stop building enormous AI data centers for money suck products with no actual net positive revenue.

Normally I try to go with the most charitable interpretation, but this article makes it difficult.

> Meanwhile China is adding power capacity at a wartime pace—coal, gas, nuclear, everything....

China is adding solar. Mostly solar. The word "solar" does not appear even once in this press release, and that seems disingenuous.

I _do_ think there's a place for more efficient use of the fossil fuels we do have. People are going to continue to burn natural gas for a while, so we might as well do it better I guess. But America isn't going to make up the energy deficit with fossil fuels, no matter how "clever" we are.

Does it make anyone a little sad that we could have actual abundance with solar and wind and nuclear?

Also, this is only commercially viable because this regime has rendered the EPA functionally powerless.

Just vomited in my mouth a little bit. A supersonic aerospace company doing a half-assed pivot into fossil fuel electricity generation to, what, try to simultaneously capitalize on AI CAPEX while also soliciting government handouts?

Come on, get serious.

AI data centers still consume a lot less than most other things on the grid. In percentages it's less than 1%. Much less. It might get to a percent in a few years. The energy demand growth from other sources is much more significant. Things like industrial heating, domestic heating and other domestic usage, transport (car and truck charging), etc. are growing much more aggressively than even the most aggressive growth scenarios for AI.

Electrification of the economy, which is a thing that at least the US is way behind on, is going to be a massive driver of electricity demand across the world. And a lot of countries are going to benefit from cost savings there. Not having to import expensive oil and gas in favor of cheaply produced solar/wind energy is going to wipe out quite a few billions from the trade balance of countries across the world. China is leading by example here. Their diesel imports are declining sharply already. Investments in renewables are rising accordingly. This is not driven by green washing but by raw economics.

For the same reason, oil and gas prices usage is predicted to enter a steady decline pretty much everywhere. The IEA (known for overly conservative oil biased predictions) is predicting this will be in decline by 2030. They are probably wrong again and it might be a few years sooner. In China next year is a better estimate.

Most growth on the grid (80-90%) is driven by renewables + battery addition to the grid. It's actually not even close in most countries. Including the US. Gas turbines are hard to get in a hurry. Most of the ones that are realistically going to be installed soonish were ordered quite some time ago. Same with nuclear reactors. Supply of those is even less elastic (decades rather than years).

In the mean time, there are hundreds of gw of clean energy (which can be ordered and brought online with very short lead times) coming online every year. Think a few dozen of nuclear reactors worth of capacity. In the US alone. Every year. Vs. a handful of nuclear reactors over the next decade. And a sprinkling of gas plants barely replacing lost capacity (closures of coal and older gas plants). All at great cost of course and typically after long delays.

A lot of the AI related fossil fuel usage growth is increasing load on existing infrastructure; which for cost reasons was being under utilized. As soon as cheaper power can be secured, that capacity will revert back to being underutilized. That's just simple economics.

Whether the US will be able to adapt to other countries doing things cheaper and better than them remains to be seen. It looks like it will have lots of expensive and obsolete gas infrastructure pretty soon. And a lot of debt that financed that. And a lot of data centers operating under high gas prices competing with data centers built close to ones with access to cheap renewables might become a thing as well. Some people are predicting a bubble. When that bursts, the more economical data centers might have a higher chance of surviving.