If the other front-age article on the Science publication about how human population will have overwhelmed the earth by now isn't enough warning, then I'm old enough to remember how the "Japanese economic miracle" was going to inevitably overwhelm the US.
There is a serious lack of wisdom in lots of these statements.
I like to remind people that multiple mainstream "AI Safety" campaigners have called for first strikes on Chinese data centers:
In a new report titled “Superpowers and AGI,” former Obama-era national security official Jacob Stokes calls for United States to begin preparing readiness drills in response to Chinese AI development — up to and including military strikes on the nation’s data centers, the South China Morning Post reported.
“Data centers can be bombed with conventional munitions,” Stokes writes. (He’s not the first thinker to suggest bombings to stop powerful AI; in 2023, AI critic Eliezer Yudkowsky suggested that countries needed to be ready to “destroy a rogue data center by airstrike.”)
> I'm old enough to remember how the "Japanese economic miracle" was going to inevitably overwhelm the US
I'm also old enough to remember that in the 1980s Japan had become the world's second-largest economy and there were serious concerns in Washington that Japan might eventually surpass the United States economically. The Reagan administration began putting considerable economic leverage and measures to create pressures intended to destabilize Japan's economy. There's historical agreement on this intent. See 1985 Plaza Accord to inflate the Japanese yen. See 1986 US-Japan Semiconductor Agreement to restrict Japan's dominance in DRAM chips.
> In a new report titled “Superpowers and AGI,” former Obama-era national security official Jacob Stokes calls for United States to begin preparing readiness drills in response to Chinese AI development — up to and including military strikes on the nation’s data centers, the South China Morning Post reported.
War Hawks gotta war hawk.
Ever since the fall of the USSR, there has been a faction of people in the USA desperately trying to invent a new "enemy" to scare everyone and drive foreign policy.
The past few years, being either compute constrained or energy constrained has driven many innovations in this space. Either way, higher availability of resources does not always equate higher utilisation.
OK, so the more energy constrained country (US) has to innovate, then china steals the innovations and leverages them even further due to their massive energy resources. In the end energy wins always.
That would more seem a side-effect of being constrained in some dimension, whether compute, energy, or else ("when does someone resort to taking something that's not theirs/ours/anyone's/etc?")
If we want to beat china at energy ASAP the only real option is tremendous amounts of solar/batteries + large reciprocating engines. A lot more of this:
Reciprocating gas engines have very fast response times. This pairs well with the potential volatility of a large solar grid. You can get extremely good performance out of these machines with the blessing of turbochargers. They sound like gas turbines when they start loading up. Kind of like a poor man's version of the real deal. It's impressive how fast you can go from zero to several megawatts: https://www.youtube.com/watch?v=SzKu9QBGI4s. The engine is at a constant RPM that entire time. It's really bananas how much damage you can do with a powerful compressor.
US people are kind of lunatically liberal in a few dimensions, like jaded person. Remember the extremely dismaying showings of political extremist behaviors?
You don't need the AI chips money to fund solar expansion. Solar panels are already tariffed exuberantly. Removing that tariff, solar expansion will not be faster at all, because there will be another random political faction coming out to eat the profit.
This country inherited a honorable tradition that is no longer fitting in this highly connected world. And few people are educated to objectively acknowledge that.
I believe the implication is that their society has different baseline energy needs before compute even comes into play. And compute requirements are also multiplicative with population.
If the trends continue it will be much quicker than 10 years. They've already done a tremendous job in Africa which is a whole other thing unto itself. If they were to focus back on their own infra with all the skilled contractors they have, they can scale as quickly as they need to.
Permitting reform is currently working its way through congress, for what it's worth.
It's federal permitting, local "NIMBY" veto control over everything, state environmental laws, tariffs on solar equipment, and out of control costs on infrastructure in the US (a combination of corruption, overpowered unions, and regulations).
Like many problems in the US, the causes are many and diffused across different layers of government, different departments, making it very difficult to solve.
Due to the laws of physics, electricity generation and usage are very nearly 1:1. Electricity storage is basically a rounding error.
I.e. if you turn on a light switch in your home, somewhere, a turbine has almost certainly just sped up very slightly with approximately a speed of light delay.
Solar panels don't run at night. Betting on batteries is betting on an exponential that's been growing at a third the speed it needs to for it to be meaningful. China's grid uses them to lower emissions during the day but has idle coal plants to keep the lights on definitely.
This is something a capitalism country can't do.
In short we need to build a five grids worth of nuclear power plants as a national security priority, the same way the interstate highway network was, if we want to have any chance of keeping up.
In the mean time we can use gas turbines to make up the short fall for the next five years.
Nukes are great, although with the impending "Idiocracy" styled future that seems to be coming I'm less keen on them as I doubt our ability to properly run them over a long timeframe.
What we need, for many reasons, is a national water grid. The grid would provide plenty of pumped hydro to store solar power and return it to the grid at night. It would also resolve flooding issues and provide food security in the face of climate change. I don't think our country thinks that big anymore though.
Geohot also says that AI will "address" the issue of how much we pay to humans vs how much energy they need... a.k.a. you should be paid just enough to eat and not die.
"knowledge workers are so grossly overpaid compared to the energy they consume"
I think he is just joking. Even though it is a bit toxic, I believe this guy is just using an exaggerated writing style to be controversial and trendy.
What a weird take. China should be generating more electricity than the US, if only because they have 4x the number of people. And regardless, "generating electricity" is not automatically some sort of advantage.
And I don't get his link to the X post that he's labeled "nationalizing nvidia"; what he links to is... not that.
I agree in general that the US is completely mismanaging the adversarial relationship with China, but this post reads like a conspiracy theorist fever dream.
Not trying to advance the Kardashev scale is silly. The US goal should be for the average middle class dude to have a 1000 humanoid robots to his name.
> Sell as many AI chips into China as possible, use the money to purchase solar panel and battery factories, and build out electricity capacity as quickly as possible.
This... isn't how capitalism works. If nVidia or whoever sells chips the money doesn't just magically go into strategic infrastructure
The future is unwritten.
If the other front-age article on the Science publication about how human population will have overwhelmed the earth by now isn't enough warning, then I'm old enough to remember how the "Japanese economic miracle" was going to inevitably overwhelm the US.
There is a serious lack of wisdom in lots of these statements.
I like to remind people that multiple mainstream "AI Safety" campaigners have called for first strikes on Chinese data centers:
In a new report titled “Superpowers and AGI,” former Obama-era national security official Jacob Stokes calls for United States to begin preparing readiness drills in response to Chinese AI development — up to and including military strikes on the nation’s data centers, the South China Morning Post reported.
“Data centers can be bombed with conventional munitions,” Stokes writes. (He’s not the first thinker to suggest bombings to stop powerful AI; in 2023, AI critic Eliezer Yudkowsky suggested that countries needed to be ready to “destroy a rogue data center by airstrike.”)
https://futurism.com/science-energy/obama-admin-official-sto...
https://www.cnas.org/publications/reports/superpowers-and-ag...
> I'm old enough to remember how the "Japanese economic miracle" was going to inevitably overwhelm the US
I'm also old enough to remember that in the 1980s Japan had become the world's second-largest economy and there were serious concerns in Washington that Japan might eventually surpass the United States economically. The Reagan administration began putting considerable economic leverage and measures to create pressures intended to destabilize Japan's economy. There's historical agreement on this intent. See 1985 Plaza Accord to inflate the Japanese yen. See 1986 US-Japan Semiconductor Agreement to restrict Japan's dominance in DRAM chips.
[dead]
> In a new report titled “Superpowers and AGI,” former Obama-era national security official Jacob Stokes calls for United States to begin preparing readiness drills in response to Chinese AI development — up to and including military strikes on the nation’s data centers, the South China Morning Post reported.
War Hawks gotta war hawk.
Ever since the fall of the USSR, there has been a faction of people in the USA desperately trying to invent a new "enemy" to scare everyone and drive foreign policy.
Yikes. If all you have is a hammer, everything looks like a nail. That's not great when your hammer is a war machine.
The past few years, being either compute constrained or energy constrained has driven many innovations in this space. Either way, higher availability of resources does not always equate higher utilisation.
The same could be said about deepseek given they had to innovate with limited compute?
Exactly, but with the miniaturisation frontier still ahead of us.
OK, so the more energy constrained country (US) has to innovate, then china steals the innovations and leverages them even further due to their massive energy resources. In the end energy wins always.
That would more seem a side-effect of being constrained in some dimension, whether compute, energy, or else ("when does someone resort to taking something that's not theirs/ours/anyone's/etc?")
If we want to beat china at energy ASAP the only real option is tremendous amounts of solar/batteries + large reciprocating engines. A lot more of this:
https://www.wusf.org/economy-business/2025-05-06/lakeland-el...
Reciprocating gas engines have very fast response times. This pairs well with the potential volatility of a large solar grid. You can get extremely good performance out of these machines with the blessing of turbochargers. They sound like gas turbines when they start loading up. Kind of like a poor man's version of the real deal. It's impressive how fast you can go from zero to several megawatts: https://www.youtube.com/watch?v=SzKu9QBGI4s. The engine is at a constant RPM that entire time. It's really bananas how much damage you can do with a powerful compressor.
US people are kind of lunatically liberal in a few dimensions, like jaded person. Remember the extremely dismaying showings of political extremist behaviors?
You don't need the AI chips money to fund solar expansion. Solar panels are already tariffed exuberantly. Removing that tariff, solar expansion will not be faster at all, because there will be another random political faction coming out to eat the profit.
This country inherited a honorable tradition that is no longer fitting in this highly connected world. And few people are educated to objectively acknowledge that.
[dead]
Of note, China has a population roughly 4x that of the US
Which advantage(s) of a larger population do you believe is involved here?
I believe the implication is that their society has different baseline energy needs before compute even comes into play. And compute requirements are also multiplicative with population.
and roughly equal (1x not 4x) land when you include hawaii and Alaska.
If the trends continue it will be much quicker than 10 years. They've already done a tremendous job in Africa which is a whole other thing unto itself. If they were to focus back on their own infra with all the skilled contractors they have, they can scale as quickly as they need to.
What’s stopping more electricity being generated in the US? It’s both financial and political, but private business could step up here?
Permitting reform is currently working its way through congress, for what it's worth.
It's federal permitting, local "NIMBY" veto control over everything, state environmental laws, tariffs on solar equipment, and out of control costs on infrastructure in the US (a combination of corruption, overpowered unions, and regulations).
Like many problems in the US, the causes are many and diffused across different layers of government, different departments, making it very difficult to solve.
regulatory capture by coal and gas companies
They're literally paying wind farm companies billions to go away - https://www.snopes.com/fact-check/trump-rwe-wind-farms/
That’s generation. What about usage? If they were at capacity, wouldn’t it be the same story?
Due to the laws of physics, electricity generation and usage are very nearly 1:1. Electricity storage is basically a rounding error.
I.e. if you turn on a light switch in your home, somewhere, a turbine has almost certainly just sped up very slightly with approximately a speed of light delay.
So, if US leads with compute power, doesn’t it keep it at an advantage?
No, some US datacenters are already at the point where they can't install more hardware due to lack of electrical power.
Why would China need AI chips from the US?
Solar panels don't run at night. Betting on batteries is betting on an exponential that's been growing at a third the speed it needs to for it to be meaningful. China's grid uses them to lower emissions during the day but has idle coal plants to keep the lights on definitely.
This is something a capitalism country can't do.
In short we need to build a five grids worth of nuclear power plants as a national security priority, the same way the interstate highway network was, if we want to have any chance of keeping up.
In the mean time we can use gas turbines to make up the short fall for the next five years.
If you were using the energy for model training, there's no reason you couldn't scale compute on solar generation.
You will have your datacenters offline at least a third of the day.
> Betting on batteries is betting on an exponential that's been growing at a third the speed it needs to for it to be meaningful.
Can you explain this statement?
> China's grid uses them to lower emissions during the day but has idle coal plants to keep the lights on definitely.
This is not what is happening.
https://ourworldindata.org/grapher/annual-co2-emissions-per-...
10 years*
Nukes are great, although with the impending "Idiocracy" styled future that seems to be coming I'm less keen on them as I doubt our ability to properly run them over a long timeframe.
What we need, for many reasons, is a national water grid. The grid would provide plenty of pumped hydro to store solar power and return it to the grid at night. It would also resolve flooding issues and provide food security in the face of climate change. I don't think our country thinks that big anymore though.
The benefits would be clear but it would also displace millions of people.
The US could have it's own Three Gorges Dam on the Mississippi River, but it displaced 2 million people.
Geohot also says that AI will "address" the issue of how much we pay to humans vs how much energy they need... a.k.a. you should be paid just enough to eat and not die.
"knowledge workers are so grossly overpaid compared to the energy they consume"
https://geohot.github.io/blog/jekyll/update/2026/06/11/ai-wi...
I think he is just joking. Even though it is a bit toxic, I believe this guy is just using an exaggerated writing style to be controversial and trendy.
Selling China chips so US can afford China's solar factories… So basically a trade deficit with extra steps?
How's the per capita energy looking? Still lower despite China being a manufacturing and industrial hub?
What a weird take. China should be generating more electricity than the US, if only because they have 4x the number of people. And regardless, "generating electricity" is not automatically some sort of advantage.
And I don't get his link to the X post that he's labeled "nationalizing nvidia"; what he links to is... not that.
I agree in general that the US is completely mismanaging the adversarial relationship with China, but this post reads like a conspiracy theorist fever dream.
Geohot, pierce through the noises as always, have the most brilliant thoughts.
he said years ago he believes factoring is in P. i wonder if he will turn out to be right.
Now do it per capita
https://ourworldindata.org/grapher/electricity-mix?tab=line&...
Not trying to advance the Kardashev scale is silly. The US goal should be for the average middle class dude to have a 1000 humanoid robots to his name.
> Sell as many AI chips into China as possible, use the money to purchase solar panel and battery factories, and build out electricity capacity as quickly as possible.
This... isn't how capitalism works. If nVidia or whoever sells chips the money doesn't just magically go into strategic infrastructure
Taxes
> If trends continue, China will have 10x the electricity of the US in 10 years.
And if trends had continued from 1900, our streets would be covered 200 feet high in horse droppings.
Oddly enough, this is a solid example of how innovation leads to improvements in infrastructure.