In the global capital market, AI is undoubtedly the rising star in recent years. However, as the development of AI technology enters different stages, the volatility of the market has become increasingly apparent.

In the latest research report released on the 10th, AI-related investments are divided into four main stages, representing different levels from infrastructure construction to productivity enhancement.

AI transactions have entered the second stage and are gradually shifting from relying solely on valuation growth to being driven by corporate profits to drive market performance.

First and second stages: Infrastructure companies take over from hardware providers

The first phase of AI transactions was dominated by hardware providers, with NVIDIA being the leading company in this stage.

Advertisement

As the market matures, investors continue to have high confidence in AI construction, and infrastructure companies begin to take over, reaching the second stage.

The focus of the second stage is on the construction of AI infrastructure, including semiconductor companies, cloud service providers, data centers, equipment manufacturers, and utility companies.

These companies benefit from the increase in AI capital expenditures, and the year-to-date gains of some infrastructure stocks have exceeded 27% in 2024. As AI technology is widely applied, the stock prices of these companies can continue to rise, and their profitability will gradually become the main driver of stock price increases, rather than relying solely on valuation expansion.In addition to this, the valuation of Phase II stocks is above average, reflecting the market's optimistic sentiment.

Compared to the equal-weighted S&P 500 stocks over the past decade, Phase II stocks are trading at a premium of 0.4 standard deviations. In contrast, the valuations of Phase III and Phase IV stocks are, respectively, 0.2 and 0.4 standard deviations cheaper.

The degree of surprise in AI spending during Phase II is diminishing, indicating that the returns on stocks in this phase may be more moderate. Nonetheless, increased demand could lead to large-scale technology companies exceeding expectations in AI-related capital expenditures.

At the beginning of 2023, the demand for Nvidia chips far exceeded analysts' expectations, and large-scale capital expenditures became increasingly positive in the first half of 2024.

However, the magnitude of surprises in Nvidia's sales and large-scale corporate capital expenditures has been narrowing. The upcoming third-quarter earnings season will provide another touchstone for AI demand and spending.

Phase III: Monetization of AI applications remains uncertain

Phase III focuses on companies that are trying to generate new revenue through AI technology, primarily software and IT service companies.

Despite the lower valuations of these companies, Goldman Sachs believes that the commercial progress of AI applications is not as ideal as expected. Although investors are full of expectations for the prospects of AI applications, the actual construction and monetization of applications still face challenges.Through an IT expenditure survey, it was found that although corporate investment in AI technology has increased in 2024, it is expected that only 3% of IT budgets will be allocated to the development and application of generative AI technology next year. This implies that AI-driven revenue growth is difficult to achieve in the short term, and investors still need to wait patiently for further development of AI technology.

However, the report also mentions that platform-type stocks may stand out in the third phase. These platform companies, including Microsoft, MongoDB, Datadog, etc., provide the best utilization methods for AI infrastructure and lay the foundation for the construction of the next generation of applications.

Phase Four: Potential Winners in Productivity Enhancement

Phase Four refers to companies that are expected to achieve productivity enhancement through AI. Goldman Sachs believes that these companies may gain the most profit enhancement in the future due to the widespread application of AI, but at present, the comprehensive popularization of AI still requires several years.

According to the survey, only 6% of companies have used generative AI in the production process, and there are significant differences across industries. For example, the application of AI in the manufacturing and technology industries is relatively widespread, while the acceptance in traditional industries is lower.

Only after the AI applications in Phase Three achieve large-scale commercial use will the companies in Phase Four truly enter the investors' field of vision. This means that although the companies in Phase Four have huge potential for growth, they are still in the short term.