The Empire Strikes Back: China Prepares One Trillion Yuan AI Plan to Rival $500 Billion US Stargate Project

[AleksandarK]

A few days ago, we reported on the US reading a massive 500 billion US Dollar package called "Stargate Project" to build AI infrastructure on American soil. However, China is also planning to stay close behind, or even overlap the US in some areas, with a one trillion Yuan "AI Industry Development Action Plan". Translating into around 138 billion US Dollars at the time of writing, the Chinese AI plan is similar to the US Stargate project: develop AI infrastructure through data center expansion and AI accelerator scale-up. Unlike the Stargate project, led by private initiatives and OpenAI at the helm, the Chinese AI Industry Development Action Plan is an entirely state-sponsored initiative that will fund firms like Baidu, ByteDance, Alibaba, and DeepSeek with additional AI accelerators (or sanction-abiding GPUs) to create more advanced AI systems.

Over the past few days, DeepSeek, a branch of a Chinese hedge fund, has not only open-sourced its R1 reasoning model but made it entirely free to use for everyone. This has challenged moat of Western competitors like OpenAI, pushing its CEO Sam Altman to offer an O3-mini reasoning model for up to 100 queries per day for the ChatGPT Plus user tier. Not only did DeepSeek provide a model equally intelligent to OpenAI's best, but it also offered it completely free. This has stirred up the tech community quite a bit and showed that Chinese companies are not much behind Western competitors. With this AI action plan, the Chinese government wants to push its domestic AI makers even further ahead and allow them to overtake cutting-edge model development potentially. Of course, getting GPUs for these projects remains an intricate task, but with export control loopholes and domestic AI accelerator development, the AI arms race is picking up quite a bit.



View at TechPowerUp Main Site | Source

 

[Daven]

The biggest obstacle to the US initiative is not money but electricity. With power generation so politized in the US, China can surpass all efforts because they will build whatever power infrastructure is required while US politicians argue over windmills causing cancer. Of course, if this power generation comes from fossil fuels, we are all screwed. Here's to hoping China will invest in wind, solar, nuclear and natural gas (still fossil fuel but a little cleaner than coal) to build out AI and leave coal in the past where it rightfully belongs.

 

[jesdals]

Ill second that @Daven

 

[ncrs]

The biggest obstacle to the US initiative is not money but electricity. With power generation so politized in the US, China can surpass all efforts because they will build whatever power infrastructure is required while US politicians argue over windmills causing cancer. Of course, if this power generation comes from fossil fuels, we are all screwed. Here's to hoping China will invest in wind, solar, nuclear and natural gas (still fossil fuel but a little cleaner than coal) to build out AI and leave coal in the past where it rightfully belongs.

In the last week the US President has signed many Executive Orders that will affect this issue:

• Putting America First In International Environmental Agreements
• Unleashing American Energy
• Declaring a National Energy Emergency
• Unleashing Alaska's Extraordinary Resource Potential
• Removing Barriers to American Leadership in Artificial Intelligence

It remains to be seen how effective will they be at attaining stated goals.

 

[Daven]

In the last week the US President has signed many Executive Orders that will affect this issue:

• Putting America First In International Environmental Agreements
• Unleashing American Energy
• Declaring a National Energy Emergency
• Unleashing Alaska's Extraordinary Resource Potential
• Removing Barriers to American Leadership in Artificial Intelligence

It remains to be seen how effective will they be at attaining stated goals.

Every one of those statements is so vague one could argue the goals can be met without doing much of anything.

• Putting America first means dropping out of every international agreement so American can just keep doing what its doing
• Unleashing American Energy means continue to generate electricity exactly as we are now
• Declaring an emergency is the same as point number one; dropping out of international agreements because foreigner's suck and are trying to hold America back
• Unleashing Alaska can never be verified that anything happened because it is so remote, no one will check
• Removing barriers means tariffs, tariffs, tariffs

Everything above is exactly what the current and past administrations have done since the beginning of time and it does not move the needle one bit in any direction. Since our government is wholly ineffectual, electricity will continue to be generated based on free market TCO. That means lots of solar and wind as they are the cheapest to produce. Since both are generated domestically using domestic parts and machinery, any administration can claim to win. Here is the forecast by the EIA for electric growth. This is the energy that will power future AI infrastructure. The only thing missing is nuclear. I'm a big nuclear proponent and this is the only place I see the government helping...by getting out of the way.

 

[TumbleGeorge]

Here's to hoping China will invest in wind, solar, nuclear

This is not hope, but facts in China. Investments in such energy sources are huge and the projects are actually happening. Around 40% of production is from non fossil sources and this percentage increase constantly.
More in Wikipedia.
Fun fact: China production of electricity is around 30% of total electricity production of the world.

 

[ncrs]

Every one of those statements is so vague one could argue the goals can be met without doing anything.

Those aren't statements, those are titles of the Executive Orders. You can read them in full at whitehouse.gov.
All depends on their execution, but don't forget that the opposition will be closely watching how all of those EOs are being implemented.

The only thing missing is nuclear. I'm a big nuclear proponent and this is the only place I see the government helping...by getting out of the way.

Nuclear is coming along as well for the AI push. Not necessarily only from the government, but private companies.

 

[Daven]

Those aren't statements, those are titles of the Executive Orders. You can read them in full at whitehouse.gov.
All depends on their execution, but don't forget that the opposition will be closely watching how all of those EOs are being implemented.

While the intent of EOs are a call for action without needing the legislation branch, they typically turn out to be no more than empty rhetoric meant to rally the base. I have only pessimism and disdain towards the US government so the only hope for increased AI infrastructure and the power needed to fuel it will mostly come from the private sector. Much like the growth of the space industry which was funded by public and private sources, I think AI and new electricity growth will be mostly reliant on investors and company private expenditures. Government funding will be ineffectively spent at best, end up in the private bank accounts of corrupt politicians at worst (probably both).

This is why IMHO China can pull this off and stay ahead of the US for a fraction of the cost.

 

[Nater]

The biggest obstacle to the US initiative is not money but electricity. With power generation so politized in the US, China can surpass all efforts because they will build whatever power infrastructure is required while US politicians argue over windmills causing cancer. Of course, if this power generation comes from fossil fuels, we are all screwed. Here's to hoping China will invest in wind, solar, nuclear and natural gas (still fossil fuel but a little cleaner than coal) to build out AI and leave coal in the past where it rightfully belongs.

And we'll never come close to competing. China builds coal plants at a rate of 2 per week. The US has maybe 2 nuclear plants under construction - total.

Wind, solar, and all the green bullshit is like pissing in the ocean with what China does, in terms of competing and "saving" the planet.

 

[ty_ger]

China builds coal plants at a rate of 2 per week.

I don't think that is true any more. It peaked in 2023. In 2024, it seems they reduced production some 80%.

 

[mb194dc]

All this infrastructure won't be needed, you can run all the models that'll be needed for a long time on what already exists...

Deepseek showing you can create a model with 5% of the resources. Such efficency gains won't end there. That's the important part, more so than the effectiveness of it.

Just massively over building and it'll end very badly economically.

 

[ty_ger]

In the last week the US President has signed many Executive Orders that will affect this issue:

• Putting America First In International Environmental Agreements
• Unleashing American Energy
• Declaring a National Energy Emergency
• Unleashing Alaska's Extraordinary Resource Potential
• Removing Barriers to American Leadership in Artificial Intelligence

It remains to be seen how effective will they be at attaining stated goals.

Just a matter of context: USA doesn't refine and use the majority of USA-collected oil. They export their crude production and refine imported crude. So, more production in US does create more money, but it doesn't reduce foreign reliance and doesn't fix potential pricing problems because they are different products and the supply and the demand of one isn't equal to the supply and the demand of the other. US refineries are mostly incapable of refining US crude, and the cost to adjust their production process is considered too great and at odds with profit motives.

 

[Daven]

If the money can be spent effectively, I think the entire $500B should be spent on making safe, cheap and small scale nuclear power plant technology that is connected directly to a data center or other power hungry facility. That is a much better investment that benefits many industries not just AI.

Such technology would also put the US ahead of all other nations at which point we could license it for huge sums of money.

 

[Prima.Vera]

China and USA are always pushing and investing more and more in IT and AI. India and Taiwan are also close by.
Meanwhile EU or Japan are just sleeping on it and losing train after train.
No wonder that both Japan's and EU's economy are going to crapper year by year compared to those of USA or China...

 

[Dr. Dro]

138 billion dollars is no small amount of money in an economy such as China's. Given the scale, if executed well, this plan could bring them much closer to catching up to, and perhaps even leapfrogging Western technology. 45+47's big red Siri/Bixby/Alexa/Grok button still hits harder, faster, stronger for the time being, but I'm unsure that will hold in the long run. I guess industrial espionage will be on the rise

 

[kondamin]

Just 183billion?

Facebook alone put 40Billion in AI last year.
Microsoft 50Billion
Amazon did 75 Billion
Google 33 billion

I have serious doubts all this money is well spent on 4bit compute

 

[Jtuck9]

While the intent of EOs are a call for action without needing the legislation branch, they typically turn out to be no more than empty rhetoric meant to rally the base. I have only pessimism and disdain towards the US government so the only hope for increased AI infrastructure and the power needed to fuel it will mostly come from the private sector. Much like the growth of the space industry which was funded by public and private sources, I think AI and new electricity growth will be mostly reliant on investors and company private expenditures. Government funding will be ineffectively spent at best, end up in the private bank accounts of corrupt politicians at worst (probably both).

This is why IMHO China can pull this off and stay ahead of the US for a fraction of the cost.

China’s Attempt to Avoid the American Tech Monopoly Trap - American Affairs Journal

One of the great paradoxes of recent economic history is how little the information technology sector has contributed to overall productivity. Economist Raicho Bojilov examined total factor productivity across the major industrial economies from the 1970s to the present and observed: Somewhat...

americanaffairsjournal.org

 

[Bomby569]

China just built a better and cheaper (by insane amounts cheaper) LLM model, they already leading, the US is done.

 

[Jtuck9]

I like the idea of each country having a citizen owned a.i. system tailored to their specific needs, like cars built for their specific road sizes (looking at you SUVs), also a big fan of Kyosho Jutaku (constrained creativity). It would just be nice if there was more global cooperation in the context of the right to self determination rather than looming mineral wars.

 

[Denver]

What you don't understand is that DeepSeek is dumber than ChatGPT4 (normal/free) most of the time.
The number of parameters doesn't mean anything.

 

[Bomby569]

What you don't understand is that DeepSeek is dumber than ChatGPT4 (normal/free) most of the time.
The number of parameters doesn't mean anything.

it's dumber in that it scored higher in every metric. That makes no sense

 

[Denver]

it's dumber in that it scored higher in every metric. That makes no sense

Test it yourself, don't rely on third party tests (You’ll soon realize that these tests barely scratch the surface.)

 

[Bomby569]

Test it yourself, don't rely on third party tests (You’ll soon realize that these tests barely scratch the surface.)

every one can get idiotic if you probe it long enough, not sure what you mean by that, i already tested it.

 

[csendesmark]

Good,
Competition is good for all of us.

 

[TechLurker]

On the power front, here's hoping SMRs pan out; there's been some plans to fast-track SMR deployments for military basing, for future datacenters, and even for fossil fuels (powering pumps, mines, and some refineries) within the next 5 years or so. Same with Hydrogen, since it can also replace coal in furnaces and power-generation too.

But China will probably successfully get their AI datacenters up and running faster than the US. They have the capability and financial power to do so.