The development of artificial intelligence is approaching a critical inflection point. According to the latest strategy report from a major brokerage, AI has nearly completed its "command-line interface" (CLI) phase. The next leap into a graphical user interface (GUI) era is set to create a strategic allocation opportunity for China's domestic computing power supply chain and Hong Kong-listed internet stocks.
The brokerage's strategy team noted in their August 9th report that a coding agent's role in AI is similar to DOS commands for early computers—powerful in capability but with a high barrier to entry, limiting its user base. Historically, the next acceleration in computer penetration relied on the 1990 launch of Windows 3.0 and its graphical interface. For AI to achieve a similar transformation, it must navigate two key hurdles: reducing the cost of semiconductor equipment and achieving the systematic integration of workflows by leading internet companies.
The report also pointed out that AI has not yet driven a broad improvement in global total factor productivity. However, China may leverage its demographic dividend of engineers to carve a unique path, reminiscent of Japan's trajectory in the 1980s. The narrowing gap between Chinese and American AI models, the rising price of domestic models against a global trend of price cuts, and the accelerating breakthroughs in China's domestic semiconductor sector all support this view. Based on this logic, the report presents a strategic bullish recommendation for Hong Kong internet stocks.
Similar Evolutionary Paths of Informatization and Intelligentization
The brokerage report argues that the evolution of AI closely mirrors that of computers, both unfolding in six stages: single function, integrated functions, continuous interaction, specialized commands, mass-market entry point, and new carriers. Currently, AI is in the "specialized commands" stage. Coding agents offer theoretically limitless capabilities, but like the command-line interface of the computer era, they suffer from high usage barriers and poor interactive experience. Users still need to interact via chat boxes, making single-task costs high and limiting widespread adoption.
The report cited a statement from the CEO of Every, Dan Shipper, who said: "We're going to run through the CLI phase quickly. The CLI is already over." The biggest difference between intelligentization and informatization lies in the pace of development: AI has completed the investment cycle in three years that took the computer era nearly a decade. This rapid expansion necessitates a new narrative to sustain AI's growth.
Two Conditions for the Move to the 'GUI Era' That Remain Unmet
Drawing on the success of Windows 3.0, the brokerage identified two prerequisites for transitioning from CLI to the GUI era, which AI currently does not fully satisfy. The first condition is hardware that offers "more power for less cost." In the 1980s, the 386 chip offered sufficient performance while its price continually declined. Memory chip prices also fell in 1990 after a brief spike in 1988, laying the hardware foundation for Windows' adoption. However, the hardware costs AI relies on are still on an upward trajectory, with storage and GPU prices rising steadily over the past few years, starkly contrasting with the 1980s' price decline cycle.
The second condition is the seamless integration of old workflows. Windows defeated Apple's Macintosh by being compatible with DOS software, allowing users to continue their existing workflows within the new graphical interface. While major internet companies are advancing local AI adaptations, such as Microsoft's Copilot and CapCut's smart editing features, the report argues these efforts remain "individual battles" and have not achieved the level of system integration seen with Windows 3.0.
Two Keys to Breaking the Deadlock: Semiconductor Equipment and Hong Kong Internet Stocks
Based on these two conditions, the report indicates that the path to solving AI's "GUI bottleneck" clearly points to two types of assets. Semiconductor equipment addresses the issue of hardware cost reduction. The report uses 1980s Japan as a reference, where large-scale fiscal stimulus drove a semiconductor industry upgrade. Within a decade, Japan significantly increased its market share in semiconductor equipment, with four Japanese companies among the world's top ten by the late 1980s, establishing a global dominance in memory chips.
The report believes that the current surge in AI investment demand, coupled with insufficient physical production capacity, is the core reason for rising hardware prices. This creates a historic window for followers. The technological breakthroughs and capacity expansion of domestic semiconductor equipment are key conditions for pushing AI into the GUI era. Leading internet companies, on the other hand, correspond to the integration of workflows. The report notes that Tencent's WeChat mini-programs already possess the core functions of most traditional software, while Alibaba has streamlined the authorization flow between payments and e-commerce. This system-level, cross-software workflow integration capability is the core of completing a Windows 3.0-like transformation and is a moat that is difficult for general companies to replicate.
Engineer Dividend: China May Forge an Independent Path
Citing total factor productivity data, the report states that AI has so far only boosted labor productivity, not total factor productivity. This is starkly different from the "informatization boom" of the late 1990s, when both factors rose in tandem. Instead, it resembles the early 1980s. The fundamental reason is that AI in the CLI era is not a universal technology; it has a significant usage barrier and has not yet led to an overall improvement in social production efficiency.
However, during the same period when U.S. total factor productivity was flat in the 1980s, Japan experienced a secondary rise in its productivity. The report attributes this partly to policy guidance, but more critically to Japan's demographic dividend of engineers at the time. In the 1970s, the number of university students rose sharply, but employment rates were low. When the personal computer era arrived, the tech industry fully unlocked the "education premium," and the employment rate for Japanese university graduates soared, driving economic prosperity and productivity improvements.
The report believes China's current situation is very similar to Japan's at that time. Data shows the gap between Chinese and American models is narrowing, overseas models are trending toward price cuts while domestic models are rising in price, and the domestic semiconductor sector is accelerating its breakthroughs. All of these are signals that China's engineer dividend is being realized through the AI era.
Strategic Bullish Outlook on Hong Kong Internet Stocks
Based on the above analysis, the brokerage offers the following allocation directions. First, the domestic computing power supply chain, including domestic chips, semiconductor equipment, and Hong Kong internet stocks, is the core theme for capturing the AI transition to the GUI era. Second, the commodity price-increase chain, which includes non-ferrous metals, coal, oil, chemicals, and new energy, offers a hard-currency attribute that can cycle through different periods. At the same time, investors should patiently wait for the gradual realization of the CPI price-increase chain, driven by balance sheet repair, such as real estate and baijiu, in the second half of the year.
Comments