Hello! I'm your financial journalist and economic expert friend. This article from "Skei Suisu Kan" is a very insightful and, frankly, somewhat critical deep-dive into the industry. It doesn't follow the majority of media's narrative that "computing power is the new oil"; instead, it takes a closer look at the "invisible costs" that are often overlooked by the general public.
To help you understand it easily, I'll first summarize the main points in plain language and then break down the article into five aspects to reveal the truth behind this global craze for computing power.
📝 Summary of the Main Points: A Global Game of "Who Pays the Bill"
The core idea of this article is that every country is racing to build AI computing power centers, but this isn't a surefire profit-making venture. It's a structural game where nations use their sovereignty, capital seeks profits, and local governments bear the costs.
In the past, the location for data centers was determined by where electricity and land were cheap. Now, building computing power centers has turned into a strategic competition at the national level. However, whether in the Middle East, Nordic countries, the United States, or western China, all regions are using their local resources (water, electricity, land, and finances) in exchange for short-term capital gains.
The harsh reality is this: The benefits of computing power are obvious and short-lived, going to multinational corporations; the costs (such as rising electricity prices, water depletion, and financial debt) are hidden and long-term, and ultimately borne by the local residents and governments. So far, no city has successfully created a closed loop that retains industries, attracts talent, and benefits local livelihoods. Everyone is just doing "computing power assembly" rather than being the "owners" of this industry.
---
🔍 In-depth Analysis: Five Easy-to-Understand Aspects
1. The Logic Has Changed: From a "Money-Saving Business" to a "National Arms Race"
How it Used to Be:
Three years ago, building data centers was simple—just find a place with low electricity costs (e.g., cheap hydropower), low temperatures (no need for air conditioning), and cheap land. Computing power was just another IT business focused on reducing costs.
How It Is Now:
Since 2024, everything has changed. Due to the Sino-US rivalry and AI technology restrictions, the world has realized that computing power is no longer just an ordinary industry; it's the "power grid" and "oil field" of the digital age, a strategic asset for nations.
It's like the competition for cheap land has turned into a competition for strategic locations. Now, countries like Norway, the Middle East, the United States, and China are all investing heavily in infrastructure, ignoring short-term returns.
The Subtext:
Previously, the winner was whoever had the lowest costs. Now, the winner is whoever has sovereignty. But this strategic investment comes with significant hidden costs, and the capital market only talks about the potential gains, never mentioning the resource depletion and financial burdens.
2. Four Types of Computing Power Hubs, Each with Their Weaknesses
The article categorizes global computing power hubs into four types, each with unique advantages and vulnerabilities:
- 🇸🇦 Middle East (Saudi Arabia/United Arab Emirates): The Wealthy but Water-Scarce Model
- Advantages: Sovereign funds, abundant land, favorable policies.
- Weaknesses: Extreme heat in the desert requires a lot of water for cooling data centers. The Middle East is short of water and relies on expensive desalination.
- Consequences: High maintenance costs; without government subsidies, these centers can't be profitable. This model depletes local water resources.
- In Simple Terms: Using a lot of money to solve a major problem, but it's borrowing from the future.
- 🇳🇴 Nordic Countries (Norway/Sweden/Denmark): The Environment-Friendly Model with Cheap Electricity
- Advantages: Abundant hydropower and low temperatures.
- Weaknesses: High electricity consumption by data centers. Denmark's grid is overwhelmed, with a quarter of its capacity used by them.
- Consequences: Governments are tightening policies and may stop providing preferential electricity rates.
- In Simple Terms: No industry can thrive if it causes power outages for residents.
- 🇺🇸 United States (Texas/Oklahoma): The Market-Driven Model with a Fragile Grid
- Advantages: High market efficiency, abundant energy, and flexible policies.
- Weaknesses: Old and fragile grids, with rising electricity prices causing public backlash (e.g., Maine has banned large data centers).
- Consequences: Capital leaves after making profits, leaving local residents with high electricity costs and grid upgrades.
- 🇨🇳 Western China (Ulanqab/Qingyang): The Scale Model with Waiting Electricity but Value Outflow
- Advantages: Supportive policies, green energy, and well-developed infrastructure.
- Weaknesses: Local resources are used, but profits flow to larger cities like Beijing, Shanghai, Guangzhou, and Shenzhen.
- Consequences: Local areas get low-end jobs and long-term energy infrastructure debts.
3. Investment Banks Warn: Computing Power Comes with Hidden Costs
Top institutions like Goldman Sachs and the BIS are raising concerns:
- Lack of Funds: Giants like Amazon and Microsoft are expected to spend over $700 billion on AI by 2026, nearly equivalent to their annual cash flow. They're borrowing heavily to fund this.
- Bubble Concerns: The BIS compares this to the internet bubble, questioning the sustainability of such rapid expansion.
- Resource Depletion: Global data center electricity consumption could double by 2026, and 40% of data centers are in water-scarce areas. The AI industry consumes 23 billion cubic meters of freshwater annually.
- In Simple Terms: The capital market only focuses on profits, not the long-term costs of resource depletion.
4. Three Levels of Profit Distribution: Who Wins and Who Loses?
This part reveals the true distribution of benefits:
- Level 1: Nations (Sure Winners): Digital sovereignty, national security, strategic assets.
- Level 2: Capital (Sure Winners): Excess profits and global benefits.
- Level 3: Local Governments (Likely Losers): Land, water, energy, financial support, and local livelihoods.
In Simple Terms: It's a pyramid structure where nations get the strategic benefits, capital gets the profits, and local governments bear the costs.
5. The Ultimate Question: Can Computing Power Become as Stable as Oil?
The article asks whether computing power can become as enduring as oil:
- Oil: Fixed in place, with industries and ecosystems tied to it. Oil cities like Houston and Daqing thrive because resources are there.
- Computing Power: Virtual, mobile, and easily relocatable. Without subsidies, companies can leave quickly.
Current Situation: No city has successfully created a closed loop that retains industries, talent, and value for local communities. Everyone is just providing computing power, with few becoming true owners.
In Simple Terms: Trends are exciting, but cities need more than just trends to thrive. The real test is whether they can sustain their resources and benefit local livelihoods.
💡 Journalist's Conclusion
This article is sobering but important. It reminds us to celebrate the potential of AI while not ignoring the costs borne by ordinary people, local governments, and the environment.
For investors and observers, this means focusing on the sustainability and local value creation of computing power centers, not just their size and popularity. Projects that rely on cheap electricity and subsidies are highly risky.
In the future, those who can solve the problem of value outflow and turn computing power into a local economic driver will be the winners in this global competition.