The artificial intelligence (AI) revolution sparked by the launch of ChatGPT has gained tremendous momentum over the last two years, with innovation continuing to ignite in unexpected waysAs the AI sector evolves, we frequently witness significant advancements, particularly in China, where domestically developed AI technologies are surprising stakeholders across the globe.

In light of this technological shift aimed at liberating human intellect, my stance is clear: the best way to hedge against the potential risk of being replaced by AI is to become a shareholder in the leading companies of this AI feastThis approach not only serves as a short-term investment strategy but also aligns with broader long-term considerationsThe globalized nature of AI requires investors to remain vigilant about emerging investment opportunities in this sector.

Among the wealth of investment options in AI, I find industry-specific exchange-traded funds (ETFs) particularly compelling

They offer diversified risk, high transparency, and appeal to both long-term investors and swing traders alikeRecently, a new entrant has joined the market, the Huabao ChiNext Artificial Intelligence ETF (159363), which tracks the ChiNext Artificial Intelligence IndexThis new ETF has increased the total number of AI indices in the market to three, alongside existing ones that focus on AI technology.

In comparison, the Huabao ETF stands out as the only one with a daily trading limit of 20%, indicating a higher level of volatility and potential for responsive investment strategiesAs we analyze the AI industry, it’s helpful to delineate the components of the AI ecosystem: the infrastructure, software, and applications.

The AI Supply Chain: Infrastructure + Software + Application

The Huabao ETF is representative of a broader trend; its component stocks span AI hardware (37.18%), AI software (32.52%), and AI applications (30.32%), effectively covering all critical segments of the AI value chain

• Tech Stocks Propel Nasdaq to New Heights
• U.S. Manufacturing Data Shows Recovery
• Ongoing Premium! These Funds are Actively Warning of Risks
• Weak Recovery of Gold Prices
• The Global Cryptocurrency Market Enters a New Phase!

The AI domain can typically be categorized into hardware, software, and application layersNotably, both the A-share and U.Smarkets showcase a fascinating interplay, often reflecting and complementing each other's hot topics within AI.

To begin with the upstream hardware sector, often referred to as the "foundational" layer of AI, we encompass computation power, data centers, and cloud servicesIn this context, the hardware industry is frequently described as one that "sells shovels"—where major corporations invest heavily in fixed assets as their initial step into AIThis mirrors developments in the domestic chip manufacturing sector, where initially equipment-based firms were the primary beneficiariesFor instance, reports suggest that ByteDance has also made significant investments in AI-related hardware recently.

NVIDIA now reigns as the global leader in AI training chips, with its GPUs setting the industry standard for large language models, positioning it as the largest beneficiary of the AI boom

Within the Chinese stock market, prominent 'hardware champions' are leading in optical module manufacturing, showcasing significant competitiveness on a global scaleCurrently, the top three holdings in the Huabao ChiNext Index are optical module companies, signifying their critical role within the AI hardware sector.

Typically, a server can consist of multiple GPUs (usually eight), which form part of larger computational clustersData centers leverage these servers to establish extensive AI networks, relying on communication devices such as optical modulesWhile the value-added optical chips remain underrepresented in local production, Chinese firms are highly competitive in the module segment, boasting five of the top ten global manufacturers like ZTE and Huajia Optical Technologies.

Next in the chain is the midstream software sector, which focuses on advanced model techniques and service provisions, epitomized by players like Google and OpenAI

In the A-share market, companies engaged in AI software development include Softcom and Inspur.

Lastly, we have the downstream application layer, which brings AI products and solutions to marketNotably, Applovin, a significant player in the AI application domain, has seen its stock price surge by over 700% this year, achieving a market valuation exceeding $10 billionThe Chinese market is also fertile ground for various application firms that contribute to this vibrant ecosystem.

Characteristics of the AI Indices:

A Strong Focus on Core AI Businesses with Heavy Investment in Optical Modules

In terms of index structure, the ChiNext Artificial Intelligence Index is notably more focused on companies whose primary business is AI compared to the earlier AI indices

It contains a higher proportion of telecommunications companies and emphasizes significant holdings in three leading optical module firmsIn summary, the index features five distinct attributes.

Firstly, the index construction method favors companies that primarily engage in AI-related activities, establishing weight rules accordinglyThe index comprises 50 constituent stocks—more than the 50 in the China Securities AI Index and 30 in the STAR Market AI IndexAdditionally, the ChiNext index has capped individual stock weightings for application firms at 2%, a measure not implemented in the other two indicesThis approach acknowledges the fact that, although these firms are involved in AI, their overall business models are often multifaceted, with a lower correlation to AI performance, thus justifying a lower cap on individual stock weightings.

Furthermore, in terms of sector allocation, the index contains a relatively higher concentration of communications-related firms

The ChiNext AI Index particularly emphasizes communications equipment and information technology services, while the STAR Market AI Index focuses more on semiconductors and integrated circuits, and the China Securities AI Index has a broader industry distribution.

Despite this, the index manages a balanced spread across the three major themes: AI hardware (37.18%), AI software (32.52%), and AI applications (30.32%). The concentration of holdings is quite notable, especially with major allocations to leading optical module manufacturers such as ZTEAs of November 30, 2024, the top ten holdings in the ChiNext Artificial Intelligence Index account for a combined weight of 46.31%, with the optical module leaders collectively constituting 19.73% of the total.

The total market capitalization and free float of the ChiNext AI Index situate it between the other two indices

Compared to the STAR Market AI Index, it boasts a higher number of constituent stocks, with a greater median and average market cap among its constituents, reflecting a predominance of mid-cap stocks (ranging from 10 billion to 50 billion yuan). A higher overall market cap facilitates amplified investment opportunities for fund products in the index.

Lastly, the historical performance of the ChiNext AI Index has been remarkableSince the beginning of 2023, it has significantly outperformed its similarly-themed counterparts, showcasing a cumulative growth of 105.41% from January 1, 2023, to November 30, 2024. In comparison, the China Securities AI Theme Index and STAR Market AI Index posted gains of 33.13% and 43.58%, respectively.

Since 2019, the ChiNext AI Index has not only emerged strongly against the broader ChiNext Composite Index but has also exhibited a considerably smaller maximum drawdown

Between January 1, 2019, and November 30, 2024, the cumulative increase for the ChiNext AI Index reached an impressive 188.41%, outrunning the returned figures of 77.85% and 100.25% for the ChiNext Index and the Growth Enterprise Market Index, respectively.

Surpassing Expectations for Domestic AI Innovation

Although the AI wave began with the large models from ChatGPT, the emergence of open-source models has facilitated a plethora of alternatives, allowing domestic AI models to develop far beyond anticipated benchmarksAccording to rankings for AI products, the top three domestically produced models as of November 2024 include ByteDance's Doubao, which ranks second globally, followed by Baidu's Wenxiaoyan and Kimi from Dark MoonAll three products boast monthly active user counts exceeding ten million.

Moreover, Doubao's recent metrics have been strikingly impressive

As of mid-December, the daily usage of tokens for its general model exceeded four trillion, with a staggering 39-fold increase in information processing call volume over the last three months alone.

The growth potential of the AI market appears boundless, as projections from CITIC Securities suggest that by 2025, the global AI chip market could nearly touch $300 billion, with the server market approaching upwards of $600 billionThe entire value chain may even exceed a trillion dollarsMarket research by other firms predicts that the global large model market could see a compound annual growth rate (CAGR) nearing 50% within the next six years, reaching an estimated size of $109.5 billion by 2028.

As we witness the waves of change brought about by this industry, it evokes a mix of fear and excitement in all of us