Central Observation | After Hengke's sharp decline, is it a bottom or a trap?

Ask AI · Will massive AI investments drag down short-term tech stock profits?

Southern Finance, 21st Century Business Herald Reporter Zhang Weize Hong Kong Report

As the US and Iran signal a easing of tensions, global markets are boosted. According to CCTV News, on March 31 local time, Iranian President Raisi stated that Iran is willing to end the war, but only if its demands are met, especially guarantees of non-aggression. Additionally, according to Xinhua News Agency, U.S. President Trump signed an executive order at the White House on March 31, indicating that the U.S. might end military operations against Iran within two to three weeks. Driven by this positive news, on April 1, the Hang Seng Tech Index opened at 4,777.67 points, up 127.85 points, a 2.75% increase.

Although today saw a rebound, looking at the bigger picture, the Hang Seng Tech Index’s performance this year has been far from optimistic, with gains previously sparked by the “DeepSeek moment” nearly fully retraced. Reflecting on 2025, thanks to breakthroughs in domestic AI large models like DeepSeek, the Hang Seng Tech Index experienced a remarkable recovery, recording a 24.85% increase for the year and briefly reaching a stage peak of 6,715 points. However, after entering 2026, this momentum abruptly halted. Under multiple negative factors, the index has oscillated downward from its high in October last year, with a cumulative decline approaching 30%, and market sentiment faces severe tests again.

On February 28, the US and Israel launched a joint military strike on Iran, prompting Iran to declare a blockade of the Strait of Hormuz, leading to systemic sell-offs in global risk assets. The Hang Seng Tech Index bore the brunt; on the first trading day after the conflict (March 2), it fell 2.89%, losing the 5,000-point mark, hitting a nearly 10-month low.

Meanwhile, the Federal Reserve’s latest stance further tightens external liquidity support. Fed Chair Powell raised the core PCE forecast to 2.7%. According to rate futures pricing after the FOMC meeting, traders have delayed the first rate cut to 2027.

Yang Delong, Chief Economist at Qianhai Kaiyuan Fund, told 21st Century Business Herald that geopolitical tensions significantly suppress investors’ risk appetite and trigger capital outflows; combined with rising oil prices and delayed Fed rate cuts, these multiple factors resonate negatively for the short-term capital markets.

The logic of this transmission chain is not complicated: soaring oil prices directly push up energy inflation, which in turn raises core inflation expectations, making the Fed more cautious about rate cuts. Unlike traditional industry indices, tech stocks’ valuations heavily depend on discounted future earnings expectations and are highly sensitive to liquidity conditions and market risk appetite—once external liquidity tightens or risk sentiment worsens, valuation compression tends to be more severe than in traditional sectors.

In the weak market where indices continue to decline, capital shows an opposite trend of accelerating inflows.

Market data shows that in March 2026, the 1-month HKD interbank offered rate (HIBOR) fell counterintuitively from 2.36% at the start of the month to around 1.95%, diverging sharply from the macro environment where the Fed maintains high interest rates of 3.50% to 3.75%.

Dongwu Securities’ Global Chief Economist Chen Lijie pointed out that this rare “high US, low HK” interest rate spread inversion is most reasonably explained by recent large-scale non-seasonal external capital inflows into Hong Kong banks, leading to a severe surplus of short-term local liquidity.

According to Wind data, as of February 25, the net subscription scale of Hang Seng Tech-related ETFs over the past half-year (August 2025 to February 2026) reached 104.73 billion yuan, with net subscriptions exceeding 34.25 billion yuan in 2026 alone, showing a clear “buy more as prices fall” left-side game.

Yang Delong believes that the contrarian accumulation of southbound funds during the index’s irrational decline fully reflects domestic institutional confidence in the long-term prospects of core stocks, indicating that the tech sector at current levels already has significant medium- and long-term investment value.

However, this wave of capital inflows also reveals the market’s complex and fragile sentiment. Guotai Securities International Securities Strategist Wu Lixian told 21st Century Business Herald that since the beginning of the year, southbound capital inflows have shown a notable shortening of operation cycles—funds tend to quickly bottom fish during sharp declines but exit swiftly during obvious rebounds, with clear trading characteristics.

Beyond geopolitical shocks, Hang Seng Tech also faces market scrutiny over earnings reports.

For example, tech giants Tencent, Meituan, and Kuaishou released their 2025 annual reports in March, but market feedback was generally below expectations. Tencent’s stock price plunged 6.8% the day after earnings; Kuaishou’s stock dropped 14% on the same day.

Wu Lixian pointed out that the overall performance of tech stocks during this earnings cycle disappointed the market, but this is understandable. He noted that Q4 last year was the slowest quarter for China’s economic growth in 2025, with overall corporate operations under pressure.

The Hang Seng Tech Index covers internet platforms, e-commerce, cloud computing, semiconductors, and consumer electronics, with Tencent, Alibaba, Meituan, Xiaomi, JD.com as core components.

Yang Delong said these tech internet companies are fundamentally still consumer-oriented, with profits highly dependent on end-user spending willingness. Currently, macro consumption growth is low, directly suppressing short-term earnings.

However, based on the disclosed 2025 annual reports, some tech giants’ fundamentals are not bad. For example, Tencent’s full-year revenue reached 751.77 billion yuan, driven by strong growth in Video Accounts and mini-games, with gross margin rising from 53% to 56%, and gross profit growth (+21%) significantly outpacing revenue growth (+14%). Meanwhile, Xiaomi’s net profit grew 44% year-on-year to 39.2 billion yuan, with both revenue and profit hitting record highs.

Yet even Tencent, with a relatively good report card, saw its stock plunge 6.8% the day after earnings release (March 19), with its market cap dropping below HKD 5 trillion—a psychological threshold.

Xu Yang, Global Partner at Tiger Securities, told 21st Century Business Herald that overall, the Hong Kong tech stocks had already priced in an “AI revaluation” earlier. When the sector had incorporated expectations of AI bringing a second growth curve before earnings, the earnings release either needed to provide stronger evidence or clearer return pathways; otherwise, a “good but not good enough” correction could occur.

Beyond earnings, the “money-burning” competition in local life services like food delivery also directly erodes profits. For example, Meituan’s performance turned from profit to loss due to fierce competition, with a full-year net loss of 23.4 billion yuan and operating losses of 17 billion yuan.

Caitong Securities estimates that in 2025, the combined “burning money” in food delivery by Meituan, JD.com, and Alibaba reached about 80 billion yuan; in Q3 alone, related investments increased by about 10 billion yuan quarter-on-quarter.

This high-cost competition explains a seemingly contradictory phenomenon within the sector. Caitong Securities noted that if we exclude Alibaba, Meituan, and JD.com—most affected by the food delivery war—the net profit growth of the Hang Seng Tech sector in 2025 would be as high as 21.7%, with overall profit expectations revised upward by 3.4%. This suggests that without this active consumption battle, the sector’s profitability outlook would be quite different.

However, policy changes may be quietly brewing a turnaround in business models. On March 30, the State Administration for Market Regulation issued a notice on further implementing the Anti-Unfair Competition Law, explicitly aiming to regulate “involution” competition and prevent “involution” in platform economies and key sectors.

Yang Delong said that if vicious price wars in food delivery and other segments subside, internet companies could focus resources on new tech tracks, potentially lifting valuation centers and profitability levels.

But substantial performance improvement still requires time. Wu Lixian offered a more pragmatic outlook, believing that regulatory improvements are a long-term positive, but given strategic considerations to capture or maintain market share, profit margins won’t recover instantly to previous high levels, and overall performance recovery will take time.

Beyond the noise of the food delivery wars, the underlying competition among tech giants in AI is quietly unfolding and becoming another key variable reshaping the industry landscape.

From the disclosed financial data, this AI arms race has seen unprecedented investment scales. Tencent’s capital expenditure in 2025 reached 79.2 billion yuan, with R&D spending hitting 85.75 billion yuan, both setting records. The report explicitly states these huge investments support the long-term development of AI businesses.

In contrast, Alibaba’s approach is more aggressive. In February 2025, Alibaba announced a three-year plan to invest 380 billion yuan in cloud and AI infrastructure, exceeding its total investment over the previous decade; its Q4 2025 quarterly report disclosed about 120 billion yuan in capital expenditure on AI and cloud infrastructure over the past four quarters.

Goldman Sachs predicts that by 2027, the combined investment of Alibaba, Tencent, ByteDance, and Baidu in AI infrastructure will reach $84 billion, a 60% increase over 2025 levels.

However, such massive investments also trigger market anxiety: are these large-scale AI capital expenditures a long-term strategic move deserving valuation premiums, or a short-term profit black hole requiring caution?

Tencent’s earnings call undoubtedly brought this market debate to the forefront. Tencent President Liu Cixin announced that AI investments in 2026 will at least double, coupled with hints of possible reductions in stock buybacks, leading to a sharp 6.8% drop in the stock price that day. This reaction reflects core investor anxiety—the commercial return timeline for AI investments is highly uncertain, yet short-term shareholder cash returns are being squeezed.

Market logic on AI is undergoing a dramatic reversal: previously, AI investments were seen as a ticket to the future, warranting valuation premiums; now, they resemble a profit-consuming “black hole,” and without clear commercialization pathways, aggressive spending triggers reverse selling.

Xu Yang pointed out that the market’s previous willingness to reward AI investments was because the first phase involved “positioning” and “narrative ownership”: whoever had models, computing power, scenarios, and user access would get valuation premiums first. Now, in the second phase, the market is asking for ROI, as the industry shifts from “proof of investment” to “proof of profitability after investment.”

Xu Yang believes that institutions generally see AI trading this year as more selective, rewarding only those companies demonstrating AI investment returns or directly benefiting from AI capital expenditure; others will be revalued downward due to high costs, profit dilution, or AI disrupting existing businesses.

Regarding this concern, Wu Lixian admitted that currently, the marginal contribution of AI models to corporate profits is not significant; their core value lies more in synergistic effects with existing business segments or in attracting traffic within ecosystems. Yang Delong also said that whether the huge costs of tech giants in AI can quickly translate into large-scale performance growth remains uncertain.

Some voices in the market believe that the current pessimistic pricing of AI investments may severely underestimate the long-term competitiveness of Chinese tech firms. Christopher Wood, Head of Global Equity Strategy at Jefferies, told 21st Century Business Herald that Chinese AI models on the OpenRouter platform have surpassed the U.S. in weekly token processing, demonstrating China’s long-term advantage in computing power coordination and energy efficiency, and this structural strength is gradually being recognized globally.

When will the market recognize corporate AI investments? Yang Delong said that if long-term AI investments can improve user experience, reduce operational costs, and incubate new business models, they could open up valuation restructuring space for the sector.

Xu Yang believes at least four key indicators need to be met: first, revenue side—AI-related revenue should sufficiently boost overall company revenue; second, profit side—see improvements in gross margins or unit costs due to AI, not just traffic; third, usage side—metrics like token consumption and paid calls are leading indicators of revenue sustainability; fourth, capital discipline—existing business profits should cover new AI investments, and profit growth should not be continuously eroded.

When will the systemic inflection point for the Hang Seng Tech Index’s recovery truly be established?

Yang Delong analyzed that ongoing geopolitical tensions, such as US-Iran conflicts, are the main reasons for recent risk aversion and capital outflows; in the future, if the Fed’s rate cut expectations are substantially revived, boosting global liquidity, it will be the most important catalyst for a comprehensive valuation recovery of Hong Kong stocks.

Additionally, Wall Street legend Michael Burry recently publicly stated that many core companies within the Hang Seng Tech sector have experienced significant stock price declines while their underlying revenue and profits have maintained steady growth, creating a historically rare valuation-disconnect with fundamentals, thus offering a unique investment opportunity.

In Wu Lixian’s view, the most powerful short-term catalyst remains the easing of geopolitical tensions; once external risks are alleviated, it will bring immediate positive effects to tech stocks and the broader market. Looking long-term, the ultimate fundamental support depends on whether China’s macro economy can sustain steady and high-quality growth.

For investors, Xu Yang believes that at this stage, it is neither suitable to heavily buy on dips based on short-term sentiment nor to be overly pessimistic due to temporary corrections. A more appropriate approach is to maintain a neutral-to-stable position: allocate to core assets with stable cash flow and profitability, while reserving some flexible positions to increase once clearer macro or industry signals emerge. The main contradiction in the current market is not deteriorating fundamentals but the downward shift of valuation centers, making position management more important than individual stock selection.

(( Intern Zhang Boyang also contributed to this article)