The stock market has demonstrated a remarkable response this year, particularly in the area of Artificial Intelligence (AI). This surge has been nothing short of phenomenal, as stocks associated with AI experience unprecedented growth, often doubling their valueSuch performance mirrors the mythical status that AI stocks seem to achieve, captivating the attention of investors and analysts alike.

However, in the ever-volatile world of stock trading, it is important to note that no stock can rise indefinitely without experiencing a correctionHistorical trends reveal that major market rallies typically follow a certain trajectoryFirst, stocks are driven up by speculation over valuations; this is followed by a period where those valuations are digested by the market; and finally, performance metrics lead to further price increases.

Since the beginning of this year, the AIGC index has risen nearly 50%, with leading stocks in this sector witnessing astonishing increases

Advertisements

History suggests that the market is now entering the valuation digestion phase; those companies able to outperform expectations will benefit, while the majority may see their stock prices drop as the market recalibrates.

Adding to the nuances of market evaluation, Huachuang Securities has introduced a method for gauging market exuberance: relative trading activity, which compares the trading volume of a sector to its market capitalizationFor example, take a market with a total valuation of 1 trillion, where AI contributes 100 billionIf the market trades at 200 billion, with AI accounting for 15 billion, then AI's relative trading activity yield will be 0.75.

Historical data suggests that when a particular sector's relative trading activity approaches its zenith, the chances for excess returns diminish significantly in the subsequent six months

Advertisements

For instance, during April to June of the previous year, the relative trading activity in the new energy sector hit an astounding 90%, only to see the overall market-reactive returns from the end of that June to December shift dismally, causing losses for investors, particularly in the sectors mentioned.

This pattern was also observable in the resounding hype surrounding the Metaverse from October 2021 to January 2022, and the so-called "Mao Index" between July 2020 and February 2021, building a consistent narrative across various speculative bubbles.

Currently, the landscape of A-shares is characterized by soaring relative trading activity levels within the AIGC and artificial intelligence indices, reaching unprecedented heightsThe buzz around AI intensified dramatically following the public debut of ChatGPT on November 30, 2022. Initially capable of performing simple tasks like drafting emails and scripts, ChatGPT quickly evolved — or rather, catapulted — into a multi-talented AI entity.

As of now, the capabilities of GPT-4 are nothing short of extraordinary: it can respond to various inquiries, generate music, manage taxes, craft poetry, and even triumph in standardized exams such as ranking in the top ten percent of the Judicial Exam, achieving a stellar 700 in SAT Math, and securing a top tier rank in Biology Olympiad competitions.

The speed of its evolution led to widespread concern among experts, including high-profile figures like Elon Musk, who, just weeks following GPT-4's launch, called for a moratorium on the training of AI models due to concerns over their potential risks to humanity.

Despite external pressures or fears, the advancement of AI technologies is bounded by the underlying technological constraints

Advertisements

Take computing power as a prime example; ChatGPT reportedly requires an astounding amount of processing capability — approximately 3640 PF-days, translating to trillions of operations every second for an uninterrupted period of 3640 daysFurthermore, as AI models continue to evolve, the requisite computational resources grow exponentiallyOpenAI's CEO Sam Altman noted that the anticipated computational capabilities needed for GPT-5 will be a hundred times the parameters of GPT-3, necessitating computational powers ranging 200-400 times that of its predecessor.

Resource limitations in computational power became evident when OpenAI temporarily halted applications for its Plus membership service, as the demand surpassed supplyPreviously, Microsoft had invested hundreds of millions to create supercomputing resources composed of thousands of NVIDIA A100 chips to support ChatGPT, yet even this investment appears insufficient amid the rapidly advancing AI landscape

• Fostering New Momentum in Biomanufacturing
• Unlocking the Value Transformation of Ecological Products
• AI Stocks Surge: Are They Overvalued?
• Enhancing Commodity Price Risk Management
• ETF Targets Zero-Day Options Investment

This suggests a critical bottleneck; the limitations of current computational capabilities impose challenges that must be overcome to sustain growth in the AI industryAddressing these constraints will require not only significant upgrades in hardware but also innovations in computational algorithms, data storage, and transfer protocols.

Looking ahead, AI is poised to lead a revolution in productivity for humanity at largeCompanies both globally and domestically are positioning themselves to benefit from this transformative waveDomestic players such as Alibaba Cloud with its Tongyi Qianwen, the SenseTime's SenseNova, Huawei's Pangu, and Tencent's Hunyuan have all made strategic forays into the AI space, demonstrating both robust research capabilities and competitiveness in the marketTheir commitment to investment and innovation in the AI sphere not only advances China’s AI industry but places Chinese companies in a formidable position on the global stage.

From an economic perspective, market competition is going to catalyze firms to ramp up their research investments, enhance their technologies, and lower costs, thereby driving the entire industry forward into a new era of advancements and innovations.

As the age-old wisdom states, people often overestimate changes within a single year while underestimating those within a decade