Just as Jen-Hsun Huang claimed that US$1 trillion will be spent on data center upgrades in the next four years, as the development of the computing power cycle enters the second stage, the dividends are gradually beginning to shift from chips to infrastructure. Among them, Morgan Stanley pointed directly at the data center, believing that it will bring a new wave of AI investment.

In the latest research report on March 1, Morgan Stanley stated that the next wave of investment will focus more on other infrastructure of AI data centers, such as servers, network equipment, cooling systems, data storage, and even physical real estate.

Semiconductors are the first to benefit, and the next step will be handed over to AI data centers

Morgan Stanley believes that the new computing cycle that appears every 10 years will increase the popularity of computing power by 10 times and follow a similar monetization path.

So as AI develops, data center investment will also go through the traditional monetization cycle from semiconductors to infrastructure to software/services.

In the past 12 months, Nvidia's stock price has risen by about 250%, and the semiconductor industry (such as the SOX index has risen by about 60%) has also led the rise, indicating that they have been the first to benefit from the initial growth of the AI industry. In comparison, other companies in the data center technology space have seen slower growth (a median of just 20%), suggesting that there is plenty of room for future investment and value growth in the data center infrastructure space.

Currently, market investment in AI data centers is accelerating, with everything from land sales to power demand already reflected.

Therefore, looking to the future, as the demand for AI data center infrastructure increases, including servers, network equipment, cooling systems, data storage, and even physical real estate, there will be a new round of investment boom.

What will constrain AI data centers?

It is worth noting that Morgan Stanley also emphasized in the research report the constraints that may be encountered in the future development of AI data centers, especially power constraints.

Compared with today, the technology and energy infrastructure required by AI data centers in the future will no longer be the same. In addition to traditional technical infrastructure, electrical and building infrastructure is also growing in importance. This is especially important as data centers are increasingly challenged by power constraints.

Relevant data shows that the power consumption of the global data center market was 10 billion W ten years ago, but now the level of 100 billion W is very common.

Although AI currently accounts for only a small fraction of the global data center size. However, according to predictions by the Uptime Institute in the United States, by 2025, the proportion of artificial intelligence business in global data center electricity consumption will soar from 2% to 10%.

Musk has also previously expressed concerns about the future of energy. At the end of last year, he said in a podcast that the United States is currently short of chips, will have a shortage of transformers in a year, and will have a power shortage in about two years.