Summary of Key Points
AI has evolved from a mere technological innovation to a pivotal factor influencing the reallocation of social resources—money, electricity, land, and labor. It is no longer confined to the tech community but has permeated various sectors such as finance, energy, and employment through mechanisms like financing expansion, energy consumption, industrial site selection, and corporate workforce adjustments. Even ordinary people are affected by this industrial cycle, as aspects like their company's budget allocation, hiring standards, and even urban energy planning are being changed by AI.
Detailed Analysis
1. Money is no longer solely coming from tech companies; financial giants are making large bets on AI infrastructure
In the past, AI investments were primarily funded by tech companies themselves (e.g., Microsoft and Alibaba using their cash flows to build data centers) or through venture capital. However, this year, the sources of funding have become more diversified and “socialized”:
- NVIDIA, in collaboration with Goldman Sachs, Blackstone, and six other financial institutions, established an AI financing platform worth over $500 billion, with funds coming from banks, insurance companies, and even pension funds. NVIDIA also guaranteed $105 billion for OpenAI’s data center in Ohio.
- Alibaba allocated HK$80 billion in new shares in Hong Kong, all dedicated to AI initiatives. American AI cloud providers issued $220 billion in debt this year (compared to only $12.5 billion last year), with overseas pension funds and insurance companies being significant investors.
In simple terms: AI is no longer just a hobby of tech companies; it has attracted investment from the entire society, much like how the internet went from being a niche interest to a widespread phenomenon.
2. AI is reshaping the allocation of electricity and land: Data centers are following the trends of energy usage
AI consumes a large amount of electricity (a large data center’s power consumption is equivalent to that of a medium-sized town). In the first seven months of this year, the electricity consumption for internet data services increased by 43.3%, far exceeding the overall societal growth rate of 4.7%. To save energy and money:
- China has implemented the “East Data, West Computing” initiative, shifting the computing needs of AI companies in the east to the west (Inner Mongolia, Ningxia), where there is more electricity and cheaper land. These eight major hubs account for 80% of the country’s intelligent computing power.
- Europe is doing the same: new data centers are being located 175 kilometers away from large cities, in areas with abundant electricity and affordable land.
In simple terms: Data centers are no longer just located where there is a talent surplus and close markets; they now need to have sufficient electricity and land resources.
3. Computing power is being built rapidly, but is it being used efficiently? The issue of underutilization leads to the concept of “computing power sharing”
Many places are building intelligent computing centers. As of March, China’s intelligent computing power reached 1882 EFLOPS, but the overall utilization rate is less than 70%, with some new centers having utilization rates below 30%. What to do?
- “Computing power banks” and “computing power supermarkets” have emerged, allowing small and medium-sized enterprises to rent idle servers and GPUs on a unified platform (for example, the Computing Power Bank of the Bank of China in Xiongan and the Computing Power Supermarkets in Hangzhou and Shanghai).
In simple terms: It’s like having an idle car at home that you rent out on a sharing platform, solving the problem of waste on one hand and the inability to afford new resources on the other.
4. Even if you don’t use AI tools, your job and company budget have already been affected
AI’s impact on ordinary people is not direct job displacement but a gradual change in corporate hiring practices and budgeting:
- Hiring: The Indian IT outsourcing industry is seeing a shift in client demands, with a focus on “fewer employees doing more work.” The demand for entry-level programmers has decreased (not due to layoffs but rather fewer new hires during expansion periods and no replacement for those who leave).
- Budgeting: Companies are no longer spending money on AI projects indiscriminately; they are more concerned with the actual benefits, such as how much AI has reduced costs or increased efficiency. A report by iResearch shows that 67% of AI projects are still in the trial and error phase, and those with unsatisfactory ROI (return on investment) will be cut.
In simple terms: Even if you don’t use AI for writing copy or making PPTs, your company may have already reallocated the budget that was intended for your department to AI tools. Or if you’re looking for an entry-level programming job, the number of available opportunities might be fewer than before.
5. The direction of AI investment is correct, but not all investments will be successful; value needs to be realized
The lessons from the internet bubble are still relevant: many internet companies failed back then, but the infrastructure they built supported subsequent growth. The same is true for AI:
- Alibaba’s AI revenue increased by 45%, but its capital expenditure increased by 75%, and its net profit decreased by 75%. Management estimates it will take three years to reach break-even.
- Upstream industries (data centers, chips) have invested heavily based on high growth expectations, but the downstream applications are still in the validation phase. If AI cannot continuously generate revenue and efficiency, some intelligent computing centers will be phased out, and the investments will be wasted.
In simple terms: AI is the future, but not all current investments will yield returns. It’s like opening a bubble tea shop; even if the direction is right, a poor location can lead to losses.
Final Conclusion
AI has transformed from a “technical toy” into a “social resource allocator.” Whether you use it or not, you are already part of this industrial cycle. The key question is whether the value created by AI will match the money, electricity, and land resources that have been invested in it.
(End of the analysis)