The recent surge in artificial intelligence (AI) has placed China at the forefront of a global technological renaissance. On February 13, 2023, the Hang Seng Tech Index illustrated this enthusiasm as it soared to new heights, reaching a level unseen in three years. At times, it even surpassed the figures recorded in February 2022, with an impressive gain of over 25% since the start of the year. Such statistics undoubtedly paint a picture of a dynamic market responding vigorously to AI advancements.

However, the excitement in the Hong Kong stock market proved to be volatile. On the same day, the Hang Seng Tech Index faced a downturn, dropping nearly 1%, while the Hang Seng Index closed down by 0.2%. Despite this, Alibaba, one of the key players in the tech arena, enjoyed a 2.55% increase in share price, reaching a peak of over 4% during the day's trading. Meanwhile, SMIC continued its upward trajectory only to end the day with a 4.07% decline. Notably, real estate stocks were among the significant losers as certain tech companies also saw their shares drop.

Experts in financial institutions have shared insights suggesting that while the AI craze may persist, the Hang Seng Index might be entering an overbought condition. Zhao Wenli, the chief strategist for Hong Kong stocks at Jianyin International, mentioned: “The emergence of DeepSeek has served as a catalyst for valuation recovery in Hong Kong tech stocks and Chinese concept stocks. Given the potential for valuation recovery, I believe the upper limit for Hang Seng Index fluctuations this year will remain around 23,000 points. However, we are currently experiencing a period of overbought conditions, particularly after a swift rise in tech stocks, leading to profit-taking pressure.”

A closer examination reveals that the current AI boom's contribution to the revenue of most companies remains limited. Despite the impressive jump of Hong Kong stocks, which have entered a technical bull market, the underlying fundamentals suggest a more cautious outlook. High-profile companies that are redirecting their resources towards AI technologies include Alibaba, where stock prices have surged nearly 50% since the year's start, and Kingsoft Cloud, whose value has approached a dramatic doubling, having risen over sevenfold since its low last November.

Since the beginning of the year, capital flowing southward into the Hong Kong market has shown a consistent trend of strong net inflows, amounting to $17 billion, with substantial investments directed towards the IT and communication services sectors. Daily net inflows have more than doubled from an average of $4.2 billion in 2024 to $8.81 billion. However, signs of significant influx from long-term overseas investors remain the missing piece in this vibrant puzzle.

After such gains, the market has succumbed to profit-taking pressures, partly due to performance not yet reflecting expectations. Wang Zonghao from UBS pointed out that the contributions of AI to the revenues of most companies are rather limited in the short term, further compounded by the heightened competition that might restrict profit growth opportunities.

For instance, Kingsoft Cloud's stock price approached 9 Hong Kong dollars, while its U.S. shares surged past $18, a stark contrast to Goldman Sachs' price target of approximately $13, which hints at nearly a 23% downside risk. Analysts are maintaining a neutral assessment of the company.

Institutional logic surrounding Kingsoft Cloud's valuation was rooted in the anticipation that DeepSeek, which predominantly relies on large-scale AI training and inference, might bolster its business due to rising demand for computing power, storage, and AI inference services. DeepSeek is one of the most-discussed large language models (LLMs) in China and necessitates vast GPU/TPU clusters for its operation, which could position Kingsoft Cloud as a key service provider.

That said, Kingsoft Cloud’s GPU resources lag behind those of industry giants like Alibaba Cloud, Tencent Cloud, and Huawei Cloud. Consequently, it becomes plausible that DeepSeek could prioritize securing resources from those larger and more powerful platforms. If Kingsoft Cloud struggles to meet the demand for high-performance compute power like H100 or A800 GPUs, it is likely to lose ground to more resourced competitors.

Currently, notable cloud platforms like Baidu, Huawei, and Tencent have publicly declared support for DeepSeek R1, offering competitive pricing structures. For instance, Baidu has set the API costs for DeepSeek R1 at 2 RMB and 8 RMB per million input and output tokens, respectively, with a free version available until February 19, 2025. The dynamics of how service providers will balance their investments against anticipated revenues remain to be seen.

On February 12, Alibaba's share price soared by 8%, propelled largely by the realization among investors that cloud services are emerging as a vital infrastructure component, with Alibaba Cloud likely to be one of the principal beneficiaries. Speculations also hint at a potential collaboration between Apple’s AI ecosystem, Apple Intelligence, and Alibaba. However, Alibaba has faced challenges due to valuation contractions over the past few years, primarily attributed to market dynamics and slowing revenue growth. Investors are keenly watching whether Alibaba's e-commerce segment can maintain momentum amidst China's ongoing economic recovery and if its various business segments might eventually spin off into separately listed entities to uplift the overall valuation.

As the market adjusts following this correction, opportunities still present themselves. Analysts suggest that while the short-term outlook indicates potential overbought conditions, the broader AI movement is still in its nascent stages, which could lead to an uptick in the future trading range of the Hong Kong stock market.

David Scutt, a senior analyst at GAIN Capital, remarked that with the Hang Seng Tech Index currently at a forward price-to-earnings ratio of just 17.4 times, it remains significantly below the average of 24.5 times observed over the past five years, suggesting that upside potential exists. Should international investors begin to rebalance their positions within the Chinese market, the ongoing bullish streak could gain more traction. Different from the short-lived spikes seen previously in September last year when the government aimed to stimulate the A-share market, the current growth seems more sustainable. Nevertheless, the RSI (14) indicator has already entered overbought territory, highlighting the vital importance of careful entry points, target pricing, and stop-loss measures. Patience in identifying optimal trading opportunities could yield a more favorable risk-return profile.

By the close of trading on February 13, the Hang Seng Index sat at 21,814.37 points. Scutt referenced key support levels at 21,728 and 21,377 points, while resistance levels loom at approximately 22,500 points, which could attract selling pressures. A breakout above this resistance would warrant attention on the double-top resistance at 23,330 points, last seen in September 2024, and further upwards towards a significant technical level of 23,870 points reached multiple times in 2021.

This assertion aligns with Zhao Wenli's observations that while foreign capital has consistently shown net inflows, the market has generally maintained a range-based short to mid-term trading strategy, with the fluctuating midpoint continuously shifting upwards from its previous levels. This change in investment style has transitioned from a focus on high-yield stocks to a greater emphasis on tech growth sectors.

He reiterated that the upper limit of the Hang Seng Index for the time being should still remain at 23,000 points, yet at present, stocks are approaching the 22,000 points vicinity and show signs of having entered overbought territory—a particular concern following a sharp rise in tech stocks, which raises questions about immediate profit-taking sentiments. The next phase will primarily hinge on the details of policy stimulus and upcoming corporate earnings reports.

According to Wang Zonghao, rebounds driven by technology nearly always see stock prices rise ahead of performance confirmations. This year, with ample liquidity and declining interest rates, AI-related equities hold tangible opportunities for valuation increases. Analyzing experiences from prior years when stocks benefitted from advancements in 4G, 5G, and cloud computing, the MSCI China Index typically saw an average hike of 50% from trough to peak, with the CSI 300 Index achieving an even higher average increase of 72%. Nevertheless, it is prudent for institutions to exercise caution in extrapolating from these historical trends, as subsequent rises in the market have mainly been spurred by macroeconomic recovery.

Wang expressed that, fundamentally, with the competitive rush among AI users (such as cloud service providers) to build ecosystems at the expense of profitability, infrastructure providers like IDC companies and hardware manufacturers will likely experience revenue contributions sooner. Nonetheless, history suggests that potential valuation expansion may be even greater, particularly for software companies.

Referencing trends from 2019-2020 (cloud computing) and 2023 (AI), he concluded that software companies typically experience the most substantial valuation increases during these periods — their price-to-sales ratios (P/S) rose by 4 to 14 percentage points. Compared to the broader market, the valuations of software shares are still down 53% from the early 2021 peak and 38% from the peak in 2023, while hardware stocks appear approximately at par with previous highs.