Summary of Key Points
The “deflationary narrative” that has seen technology products becoming cheaper over the past few decades is being challenged by AI: The surge in demand for AI-driven hardware has led to soaring prices of chips, storage, and other technology products, which not only contributes to higher core inflation in the United States but may also raise the global inflation rate from the traditional 2% to around 3%. In the short term, this technological inflationary pressure is expected to persist for at least two years. Although AI has the potential to boost productivity and alleviate inflation in the long run, prices are unlikely to return to previous deflationary levels; the inflation rate will slow down but not reverse.
1. Has the era of technological deflation truly ended?
Over the past 25 years, the prices of technology products (such as computer software and accessories) have been declining at a rate of 5.3% annually. For example, the phone or computer you bought this year might be cheaper next year. However, things have changed:
- In July, the prices of information technology goods in the United States rose by 1.4% month-on-month, which is seven times the increase in core goods;
- The prices of video and information processing equipment rose by 12.2% year-on-year, contributing 0.4 percentage points to core inflation (similar to the impact of tariffs);
- Since November last year, the prices of “computer software and accessories” have been rising at an annual rate of 73%, accounting for more than half of the core inflation increase during that period. In short, while technology products used to save you money, they now tend to cost more.
2. How does AI drive up inflation?
AI doesn’t directly create money, but it raises prices through three main channels:
1. Hardware demand surge: Large companies (such as Amazon and Google) are buying large quantities of GPUs (graphics processing units) and storage devices to build AI data centers, leaving ordinary consumers and small businesses out of the market. As a result, prices have increased. For instance, the cost of storage components has risen from 15%-18% to 35%, and products from companies like Apple, Lenovo, and Microsoft have seen price increases of 10%-25%.
2. Absorption of energy and minerals: AI data centers consume a lot of electricity. Goldman Sachs predicts that consumer electricity prices in the United States will remain at 6% between 2026 and 2027. Additionally, the demand for copper used in data centers is equivalent to the annual output of a medium-sized copper mine, and the demand for gallium is expected to account for 10% of global demand by 2030. These resources are already in short supply (also needed for renewable energy and military applications), so their prices have risen significantly.
3. Increasing costs of software services: Subscription fees for generative AI tools like ChatGPT Plus are getting more expensive, forcing both businesses and individuals to pay more.
3. How long will short-term inflationary pressure last?
Experts generally agree that it will last for at least the next two years:
- Supply-side improvements are slow. The production of storage chips and semiconductors is not expected to increase until late 2028 to mid-2029. Unless technology giants suddenly stop spending, hardware prices are unlikely to fall significantly.
- This factor is more persistent than other inflationary drivers. Compared to tariffs and rising energy prices in the Middle East, AI infrastructure is a long-term investment, and the demand for related products will not disappear suddenly, meaning technological inflationary pressure will continue.
4. Can AI alleviate inflation in the long run?
Yes, but prices will not return to previous deflationary levels:
- In the long term, AI and robots will replace human labor, increasing productivity and reducing labor costs, which can help mitigate wage inflation.
- However, prices will not return to previous levels. Even if technology giants cut spending, hardware prices are unlikely to drop significantly; at most, inflation will slow down (for example, from 5% to 2%) rather than falling into deflation.
- The new inflationary trend: Inflation is likely to stabilize around 3% over the next decade, replacing the traditional 2% due to persistent factors such as energy shortages, competition for raw materials, and demand from AI.
5. What should individuals do?
1. Buying technology products (such as computers and phones) may cost more, especially high-end models.
2. Electricity and internet costs may increase as data centers consume more resources.
3. In the long run, inflation will not be as low as it used to be, and the returns on savings may not keep up with inflation. It’s advisable to consider inflation-resistant investments (such as stocks and gold).
In summary, while the benefits of AI are undeniable, it also means higher prices for consumers. In the short term, we need to adapt to rising costs, and in the long run, we should prepare for a higher inflationary environment.