When AI Is No Longer Just “Code”: Understanding This Counterintuitive Transformation Affecting Power, Work, and the Global Landscape
Hello everyone, I’m your financial observer.
At the recent Shanghai Bund Conference, Yang Yudong, the chief editor of Yicai, presented a set of quite counterintuitive views. When most people think of AI, they think of large models, chips, algorithms, and computing power. However, Yang Yudong pointed out that focusing solely on these aspects might lead us in the wrong direction.
The core of this article is to remind us that AI is no longer just a “future technology” in the lab; it has become a “foundation” in our reality. Just as when we discussed the internet, people focused on web pages and apps, but it was the underlying infrastructure—fiber optics, servers, and data centers—that really drove profits and changed the world. The same is true for AI; its underlying logic is undergoing significant changes.
To make this easier to understand, I’ve broken down the article into five key points and explained them in plain language.
---
1. Stop Focusing on Chips; the Real Bottleneck Is “Electricity”
Key Point: In the past, people thought the challenge with AI was its “brain” (the algorithms), but now it’s its “energy needs” (electricity).
Plain Language: Imagine you have a super smart dog (a large AI model) that reacts quickly and can do everything. But if there’s a power outage or not enough food, no matter how smart it is, it can’t function.
This is the current bottleneck in the development of AI worldwide.
- Past Misconception: Everyone was competing to see whose chips were more advanced or whose models had more parameters.
- Current Reality: The more computing power you have, the more electricity it consumes. A large AI data center can use as much electricity as a small city. Therefore, the most certain investment opportunity in the next decade won’t be in developing more complex algorithms but in power generation, transmission, and storage.
- Why This Matters: No matter how AI technology evolves, as long as it’s running, it needs electricity. This is a stable industry. Companies that produce transformers, high-voltage wires, or even small nuclear reactors (SMRs) are becoming the new pillars of AI infrastructure. It’s like building highways; no matter how fast the cars (AI applications) are, if the roads (the power grid) aren’t good, it doesn’t matter.
---
2. AI Won’t Make You Unemployed, but It Will Take Away the “Boring” Parts of Your Job
Key Point: AI doesn’t replace your job; it takes away the repetitive, tedious, and rule-based tasks.
Plain Language: Many people worry, “Will AI make me unemployed?” Don’t panic, but you need to adapt.
What AI does best is handle repetitive, boring, and well-defined tasks, such as:
- Accounting (AI is faster and more accurate);
- Customer service (AI can answer common questions tirelessly);
- Translating documents (AI provides results instantly).
These are often the most disliked tasks because they are energy-consuming and don’t generate much creative value. Once AI takes over these tasks, the nature of your job changes.
- Previously: You were the executor, responsible for completing the tasks.
- In the Future: You become the judge and creator. You’ll need to tell AI what to do, check its accuracy, and handle tasks that require human judgment and creativity.
Conclusion: The future core competitiveness won’t be about being faster than AI but about understanding human behavior better and making more informed decisions. You need to shift from being a worker to someone who directs AI.
---
3. China’s Hidden Ace in AI: The Power Grid and “Dimensional Reduction”
Key Point: When Chinese AI goes global, it’s not because of its cutting-edge technology but because of its practicality and affordability. China’s strength lies in its robust power system.
Plain Language: Here are two practical observations:
1. Regarding Chinese AI’s Global Expansion: Silicon Valley companies still strive for perfection in technology, while Chinese companies focus on quick implementation and cost control. Think of smartphones: Apple aims for the best experience, but Chinese brands like Xiaomi and Huawei quickly made technology accessible and affordable globally. In developing countries like Southeast Asia, the Middle East, and Latin America, what’s needed is AI that’s easy to use, affordable, and quickly deployable. Chinese companies excel in this combination of practicality and cost-effectiveness.
2. China’s Power Advantage: As mentioned earlier, the bottleneck for AI is electricity. China has:
- The world’s largest power grid;
- Vast renewable energy installations (wind, solar);
- Advanced high-voltage transmission technology;
- Complete power equipment manufacturing capabilities.
This gives China a natural advantage in the global AI race, especially in terms of power infrastructure.
---
4. Open Source Models Make Technological Barriers Difficult to Maintain
Key Point: Want to monopolize AI technology? The open-source ecosystem makes it hard to do so.
Plain Language: In the past, tech companies kept their technology private, requiring users to pay or use their ecosystem. Now, open source has become the norm. Many powerful AI models have open-source code that anyone can use, modify, and optimize.
This leads to two outcomes:
1. Barriers Become Ineffective: It’s hard to maintain a monopoly by preventing others from accessing the technology, as it’s already widely distributed.
2. Competition Shifts: Since everyone can access similar basic models, the competition shifts to application scenarios and data accumulation: Who understands healthcare better? Who has more data? Who has more innovative business models? These are the real barriers.
---
5. Invest in New Directions: Don’t Just Follow the “Hot” AI Companies
Key Point: The truly profitable investments may be in less glamorous, “invisible champions.”
Plain Language: In the stock market, people often follow hot trends like AI stocks. But Yang Yudong suggests three areas that might be underestimated:
1. AI Energy Infrastructure: Electricity is a bottleneck. Companies that develop liquid cooling technology, UPS power supplies, and efficient power chips are in high demand. This is a trillion-dollar market with high certainty.
2. AI-Driven Cross-Border Investments: AI has changed the logic of global investment. Instead of focusing on labor costs, investors now look at data localization and computing power infrastructure. There are opportunities in exploiting differences in AI acceptance, regulations, and infrastructure across countries.
3. “Second-Stage Chain” Hidden Champions: These companies are often overlooked. While the first-stage companies (chip and model developers) are well-known, the second-stage companies that help AI integrate into traditional industries (e.g., installing quality inspection systems in factories, integrating diagnostic data in hospitals, optimizing logistics algorithms) are less visible but are essential for the long-term success of AI.
---
Conclusion: A Spiral Upward, Not a One-Time Replacement
Finally, Yang Yudong emphasizes an important mindset: AI brings a continuous “spiraling upward” rather than a one-time replacement.
- Old jobs are replaced, but new ones are created (e.g., AI prompt engineers, AI ethics consultants).
- Old skills become obsolete, but new abilities are developed (e.g., human-computer collaboration).
- Old business models disappear, but new value networks emerge.
Tips for Everyone:
1. Don’t panic about unemployment, but be concerned about outdated skills. Learn how to collaborate with AI and use it as a tool, not a competitor.
2. Focus on Underlying Opportunities: As an investor, look at areas like power, infrastructure, and the transformation of traditional industries.
3. Understand the Power of Open Source: Don’t rely on technology monopolies; focus on companies that excel in application development and have unique data.
In the AI era, the winners won’t be those who understand code the best but those who know how to use AI to solve real problems.