In my previous article about edge AI, I discussed the business perspective of Total Cost of Ownership (TCO). That's an important angle, but not the whole picture.
These past few days, international tensions have flared up again. The situation involving Iran and the United States—something that would have been shocking five years ago—now feels routine. I started thinking about another question: in an increasingly turbulent world, what does AI really mean?
After thinking about it for a long time, my conclusion is: AI is not just a business issue, but a power issue.
Computing Power is Power
AI is changing the fundamental formula of economic operation.
Past: Capital + Labor = Productivity
Now: Capital → Computing Power + Electricity + Models = Productivity
The essence of this change is that capital can be directly converted into productive forces, bypassing human labor. Moreover, as embodied intelligence and robotics develop, this conversion efficiency will only increase.
And productive forces have always equaled power.
In feudal times, whoever owned more land and slaves had productive forces and was the ruler. In the industrial era, capital became the source of power, but still required labor, so the concentration of power was limited. Now, if computing power can directly become productive forces, and robots can replace human labor—then a tiny minority controlling immense productive forces is no longer science fiction.
The Trap of the Cloud
Imagine a scenario.
Over the past few years, an ordinary person or a small business has grown accustomed to obtaining AI capabilities through cloud services. Their work, life, and production completely depend on intelligent services provided by a few large platforms.
What happens then?
The most direct consequence is price hikes. Cloud service providers can slowly raise prices to extract surplus value from users. Just like landlords raising rent—if you can't pay, you have to leave. But the problem is you can't leave—your entire workflow is tied to this platform, and the migration costs are too high to bear.
Going further, if conflicts escalate, cloud providers can directly cut off your access. Or more subtly, limit your API call quotas or degrade your service quality. You have no bargaining power because you hold no chips.
But these aren't my biggest concerns.
In the past, when a group of people united, they could generate great power because humans were the core of productive forces. Capitalists needed workers; landlords needed peasants. But in this new era, if a small minority can possess immense productive forces through AI and robots, do they really still need so many people?
I don't know the answer. But this question is worth serious consideration.
The Significance of Edge Computing
This is where edge AI devices derive their value. It's not a cost optimization solution, but a power counterbalance.
Edge devices have three characteristics: cheap, distributed, and difficult to control.
NVIDIA's DGX Spark costs over 30,000 RMB, AMD's Ryzen AI Max+ 395 costs less than 20,000 RMB, and China's E300 module can run 32B models with a volume of less than 10 cubic centimeters. They don't need to be concentrated in a data center; they can be distributed across millions of households. And once you own the device, it's yours—no one can remotely shut it down.
It's like in a country where one side owns aircraft and artillery, while the other side is just a group of civilians with rifles. Aircraft and artillery are indeed powerful, but limited in number. The rifles in civilians' hands may be weak in individual combat power, but they win through sheer numbers and wide distribution. The two can form a counterbalance.
Kevin Kelly, founder of Wired, said in 2025: Currently, over 70% of investment is concentrated in centralized cloud computing, but actually 70% of computing already happens on edge devices. The future AI architecture will likely be hybrid—the cloud handles training, the edge handles inference, and dominance will gradually shift toward the edge.
The Technology is Ready
Edge AI is not a distant future, but a present reality.
30B parameter models already run well on edge devices. WeChat Reading runs a 30B model on the Snapdragon 8 Gen4's NPU with only 1.4 watts of power consumption. Houmo AI's chips can run 70B parameter models at 10 watts. Edge devices capable of running 30B models have dropped in price to around 10,000–20,000 RMB. Compared to annual API costs of tens or hundreds of thousands in the cloud, the payback period for this investment is short.
Electricity is the main cost, but edge devices are far more energy-efficient than the cloud—ARM processors can handle the same tasks with power consumption 10,000 times lower than the cloud.
The open-source community also provides a large number of high-quality models: Llama, Qwen, DeepSeek... You don't need to depend on any single company.
The technical barrier is rapidly lowering. In 2026, it will be entirely realistic for an ordinary person to own an edge device capable of running 30B models.
From Development to Survival
In recent years, we've grown accustomed to viewing problems through the lens of "development": how to grow, how to make money, how to scale.
But the world is changing. The international order is collapsing—the phrase "collapse of rites and music" (li beng yue huai) is no exaggeration. In this environment, the perspective for thinking about problems needs to shift from "development" to "survival."
Survival means you cannot completely rely on external supply; you need your own backup plans; you need to possess things that others cannot take away.
Edge AI devices are such things. They give you options when cloud services raise prices, a retreat when platforms cut off supply, and autonomy when the external environment changes dramatically.
This isn't just an issue for individuals or businesses. If all intelligence in a society is concentrated on a few platforms, that society is extremely fragile. Distributed, decentralized edge AI can enhance the entire society's resilience.
Final Thoughts
I'm writing this article not to promote how great edge AI devices are—the business case was covered in my previous article.
What I want to say is: don't view AI merely as a business tool. AI is becoming the core of productive forces. Whoever controls AI controls productive forces, and thus holds power.
In this sense, the value of edge AI devices far exceeds their price tag. They allow ordinary people to possess certain intelligent production capabilities, rather than completely depending on "charity" from the cloud.
In 2026, the world is becoming increasingly turbulent. In this environment, owning things you can control is rational.
Edge AI is such a thing.
Originally published at https://guanjiawei.ai/en/blog/edge-ai-power-structure
Top comments (0)