A report from China Securities Co., Ltd. indicates that in the first half of 2026, the primary global macroeconomic theme is dominated by supply shocks triggered by the U.S.-Iran conflict, with markets transitioning from "trading stagflation" to "pricing recession." U.S. stocks are navigating between the AI narrative and stagflation risks, with structural trends expected to outperform broad market trends. For Hong Kong stocks, the Hang Seng Tech Index may subsequently outperform the Hang Seng Index, having undergone more substantial adjustments and still positioned to benefit from its AI option value before a pro-cyclical market emerges.
Reflecting on the two main investment themes in the internet sector for 2025, one involved companies like Alibaba, Kuaishou, and Baidu individually showcasing AI logic, while the other focused on high-quality vertical sectors such as health, music, and mobility. Looking ahead to 2026, regarding AI, both potential returns and probability of success are currently attractive. For high-quality verticals, considering factors like growth potential & valuation levels, competitive landscape & immunity to AI substitution, and regulatory stance, the institution believes leading players in sectors like second-hand e-commerce, pharmaceutical e-commerce, industrial goods e-commerce, Online Travel Agencies (OTA), and mobility still hold significant investment value.
Key views from China Securities are as follows:
**Domestic Macroeconomic Outlook: Structural Reshaping Under Policy Support.** Domestic economic growth momentum stems from resilient external demand and a structural recovery in domestic demand. The economy had a strong start in the first quarter, with exports performing robustly due to "front-loading" effects and advantages of a complete industrial chain. Consumption showed marginal improvement driven by policies like the "replace old with new" scheme, but endogenous momentum still requires bolstering. Investment saw infrastructure efforts, steady manufacturing investment, while the real estate sector remains in a deep adjustment phase, though the rate of decline is narrowing. The full-year GDP growth target is expected to be in the range of 4.5%-5%, providing room for structural adjustments. Monetary policy continues its "appropriately accommodative" stance, with the central bank committed to maintaining stable financial market operations. Policy tools (such as RRR cuts and interest rate cuts) are ample, but their deployment depends on specific triggers, demonstrating strong strategic resolve. On fiscal policy, fiscal and monetary policies will work in synergy, with stabilizing growth as the main theme. The deficit-to-GDP ratio is maintained around 4%, with tools like ultra-long-term special government bonds and special-purpose bonds continuing to exert force. Monetary policy will coordinate with fiscal policy to jointly cushion external shocks and support the domestic economy.
**U.S. Macroeconomic Outlook: "Weak but Not Broken" Under the Shadow of Stagflation.** The U.S. economy exhibits characteristics of being "weak but not broken," supporting the US dollar, but the labor market is showing signs of fatigue. Non-farm payroll growth has slowed significantly, suggesting the economy may be entering the final stage of a K-shaped recovery. With the 2026 midterm elections approaching, the political cycle significantly disturbs economic policy. Regarding inflation, geopolitical conflicts are driving up costs and raising the inflation baseline. The U.S.-Iran conflict caused a sharp rise in Brent crude prices, triggering cost-push inflation and complicating the U.S. inflation problem. While core inflation has moderated, it remains sticky, and energy price shocks could interrupt the downward trend, forming a "stagflation-like" scenario. Influenced by geopolitical conflict and rebounding inflation, market expectations for rapid, substantial Fed rate cuts have been significantly revised. The Fed maintained the federal funds rate at 3.50%–3.75% in its March meeting. The Fed is caught in a dilemma between "controlling inflation" and "stabilizing growth," with a policy stance leaning hawkish. On the other hand, AI's impact on the U.S. macroeconomy presents a double-edged sword, boosting productivity but causing structural unemployment. While enhancing total factor productivity supports growth and profits, forming the core narrative for high U.S. stock valuations, it is also impacting the job market, evolving from hiring slowdowns to layoff risks, thereby suppressing aggregate demand and exacerbating economic downturn risks. In this "stagflation-like" macro environment, global equity markets face overall pressure. U.S. stocks are caught between the AI narrative and stagflation risks, with significant increases in individual stock volatility, making structural trends preferable to broad market trends.
**Hong Kong Stock Market: Hang Seng Tech May Outperform Hang Seng Index, Nearing a Left-Side Accumulation Period Amid Valuation Digestion and AI Narrative Restructuring.** Since the beginning of the year, Hong Kong stocks have been affected by weak domestic demand coupled with external shocks, with the Hang Seng Tech Index significantly underperforming the Hang Seng Index. This divergence is disconnected from the AI industry trend, primarily due to two reasons: first, most domestic AI progress since the start of the year has been unrelated to heavyweights like Alibaba and Tencent in the Hang Seng Tech Index; second, the composition of the Hang Seng Tech Index is more consumer-oriented than tech-oriented, continuously pressured by insufficient domestic demand and intense internal competition, with the internet sector additionally bearing the dual impact of regulatory scrutiny and AI disruption fears. The internet sector currently faces two core headwinds: first, a temporal mismatch between AI investment and profit realization, with major players significantly expanding capital expenditure but commercial transformation remaining uncertain; second, intense internal competition, exemplified by the food delivery wars, continues to suppress profit release and valuation recovery for leaders. After sufficient adjustment, the valuation attractiveness of the Hang Seng Tech Index has become significantly prominent, ranking among the lowest valuation percentiles of major global broad-based tech indices, providing a high margin of safety for medium to long-term investors. A systemic recovery for Hang Seng Tech requires the convergence of three signals: "liquidity inflection point, AI commercialization verification, and eased competitive landscape." However, its long-term upward core logic is supported by being at the starting point of the AI technology cycle, excellent organizational structure and talent advantages, and the high elasticity of the TMT sector combined with the alpha attributes of major consumption. Medium to long-term, as overseas economies move further into recession, domestic demand-side policies are expected to intensify under external demand pressure. As a core pro-cyclical asset, Hang Seng Tech is poised to benefit from a systemic revaluation catalyzed by policy.
**E-commerce & Local Life: Competition Expected to Gradually Rationalize.** In response to the escalating "internal competition" among food delivery platforms since 2025, regulators have not adopted a one-size-fits-all强硬 approach but have implemented precise window guidance through a combination of measures including multiple rounds of talks, issuing national standards, and organizing administrative guidance meetings. Entering 2026, normal market competition behavior is expected to continue in the short term. Further easing of the industry landscape in the second half of the year requires the convergence of internal and external variables: relying both on ongoing regulatory normalization and platforms proactively reducing costs and improving efficiency, and more importantly, waiting for increased AI investment to translate into cash flow pressure for Alibaba affecting its food delivery business. Short-term, the food delivery industry remains in a transition period from "competition bottoming out — profit recovery," with the second quarter still needing attention on 618 promotional investments. Taobao Flash has shifted its focus this year to half-day delivery, potentially concentrating on high-visibility paths to reduce losses and traffic conversion in the second half. In far-field e-commerce, the limited scale of near-field instant retail will not alter the broader trend of easing competition in traditional e-commerce. The growth rate differential within the industry is gradually narrowing, with second-order derivatives continuously improving, potentially leading to an inflection point in first-order derivative growth within the next 1-2 years. Under存量 competition, traffic is becoming more rigid, while the continuous increase in the online penetration rate of goods supports merchant demand, and industry monetization rates are entering an upward trajectory. E-commerce growth in 2026 faces temporary pressure due to the high base effect from previous national subsidies and a slow consumption recovery, with new policies like the "ad tax" and value-added tax reforms bringing structural impacts. The empowering effect of AI on e-commerce platforms' core businesses will gradually become apparent. In the coming phase, some companies are likely to see AI optimize and transform their traditional e-commerce operations, thereby further supporting valuation expansion for their core businesses.
**Autonomous Driving: Robotaxi Scaling, LiDAR Standardization, and Chip Architecture Reshaping.** 2026 is a critical inflection point for China's autonomous driving industry, shifting from "feature competition" to "institutionalized commercial deployment." During 2024-2025, the industry focused primarily on the rapid penetration of L2+ advanced driver-assistance systems (ADAS), the expansion of Navigate on Autopilot (NOA) capabilities, and competition among automakers for ADAS parity. In 2026, L2 enters a period of stronger regulation, while L3 moves from pilot programs to limited commercialization, with liability boundaries, access mechanisms, and standard systems beginning to take shape. L4/Robotaxi enters a phase of scaled verification, driven by simultaneous developments in China and the U.S. Based on this industry inflection point, sub-sectors like robotaxi, autonomous driving chips, and LiDAR are poised to benefit significantly. 1) Robotaxi: 2026 marks the inflection point for the industry transitioning from regional pilots to scaled commercial deployment, primarily due to: firstly, conditions for scaled deployment on the supply side have matured significantly; secondly, unit economics have shown substantial improvement; thirdly, declining cost curves and improving operational efficiency are creating synergy. Following a systemic failure incident involving Wuhan's "Radish Run" on March 31, 2026, regulatory requirements for technical and operational thresholds may increase. On the other hand, overseas expansion opens a second growth curve for robotaxi. The significance of international markets for leading Chinese L4 companies lies not only in increasing fleet size and city coverage but also in providing better unit economics and longer profit curves. 2) Autonomous Driving Chips: Industry trends for 2026 show two core directions: on one hand, integrated hardware-software solutions are beginning to demonstrate comprehensive advantages, representing the most significant industry change (rather than pure computing power upgrades); on the other hand, the evolution path for cockpit-domain fusion technology is clear and beginning to see mass production realization. 3) LiDAR: L3 represents a major opportunity for LiDAR. While LiDAR might not be the ultimate technical solution, from a practical industry perspective for 2026-2028, L3 is the most certain source of incremental demand for LiDAR. Looking at 2026, volume-wise, "standardization" of LiDAR is driving penetration rates higher, while "multiple units per vehicle" is becoming a more important growth factor. Costs have shifted from "precipitous decline" to "approaching the bottom," indicating the industry has entered a healthy development stage of "stable prices, rising volumes." From an industry trend perspective, the most significant trend in LiDAR over the past half-year is the elevation through self-developed chips, with technological paths evolving towards digitization and chip integration.
**Gaming: Maintaining an Upbeat Industry Outlook, with Concentration at the Top and AI Empowerment as Core Investment Themes.** China's game market actual sales revenue reached 350.789 billion yuan in 2025 (year-on-year +7.68%), with user scale exceeding 683 million. However, the growth driver has shifted from "user红利" to "ARPU提升," with structural differentiation deepening: self-developed revenue growth (+11.64%) significantly outpaced the overall market, with top-tier evergreen IPs and hit new titles contributing the majority of the增量, intensifying survival pressure for mid- and long-tail products. Mini-program games emerged strongly, with revenue reaching 53.535 billion yuan in 2025, becoming the largest structural增量. Client games benefited from top-tier evergreen products and simultaneous PC releases of popular mobile games, recording a significant year-on-year revenue increase of 14.97%. AI integration capability is becoming a new moat for top-tier concentration, with approximately 60% of Chinese game studios having incorporated generative AI into their development processes. Leading companies are further widening the efficiency gap with mid-tier players due to advantages formed by AI infrastructure investments. Looking ahead to 2026, the simultaneous drivers of continued license supply, the AI efficiency revolution, and cultural breakthrough through global expansion are expected to sustain overall industry景气度, but competitive intensity will also rise. Concentration at the top is the most certain trend, suggesting a focus on leading developers possessing AI R&D barriers and global capabilities.
**Entertainment: Short-form video, centered on AI video models, transitions from "technology competition" to the "first year of commercialization"; Soda Music leads to renewed differentiation in the music industry landscape, awaiting a competitive inflection point.** 2026 marks the first year of AI video commercialization. Kuaishou's Kolors has already demonstrated revenue generation, with projected Annual Recurring Revenue (ARR) exceeding $300 million in 2026. ByteDance's Seedance 2.0 holds a stronger position within the production-distribution闭环 of Douyin, Jianying, and Huoshan; priority should be given to monitoring progress in advertising, short dramas, and AI comic drama落地. AI video generation represents a major industry trend. Current AI penetration remains very low (single-digit percentages). Compared to competition, the progress of industry-wide penetration提升 is more crucial; it is far from the stage of dividing the pie. From this perspective, Seedance is enabling top players to collectively expand the market size. Secondly, based on differing model capabilities and company strategies, there is错位竞争 in terms of market, user base, and ecosystem scenarios, suggesting a non-zero-sum game. The institution believes that after a long development cycle involving rampant piracy, legalization, exclusive copyrights, and non-exclusivity, China's online music industry is already a mature sector with high competitive barriers. Although Soda Music has shown impressive growth backed by Douyin's strong traffic, it has not constituted a so-called "dimensionality reduction strike" on the existing copyright and付费 barriers of the industry. Instead, it has found its own niche within the music ecosystem. Therefore, the rise of Soda Music is seen not as directly competing for existing users but rather as promoting structural stratification within the music market and further expanding the overall user base.
**U.S. Tech Stocks: Narrative Shifts from "Faith Premium" to "Commercialization Realization," Intensifying Internal Divergence; Focus on Distribution Moats and Capital Expenditure Return Certainty.** Entering 2026, investors are no longer asking "Do you have AI?" but rather "How much revenue does your AI generate?" The stock price correlation among major AI companies has plummeted from 80% to 20%, signaling the market is strictly differentiating between beneficiaries and losers of different AI logics. The sector currently faces triple structural pressures: firstly, the core capabilities of multiple high-investment pre-trained models are converging, creating industry-level redundant sunk costs; secondly, DeepSeek, with training costs around $5.6 million, has颠覆 the logic that "high investment = high moat," proving that compression efficiency itself can be颠覆; thirdly, the AI capital expenditure race has entered a "no exit" prisoner's dilemma, with projected 2026 CapEx for the top five hyperscalers上调 to approximately $750 billion, while AI services generated only about $25 billion in revenue in 2025, making the ROI realization timeline the most critical variable in valuation frameworks. On an individual stock level, Google has重置 market perception with its Gemini 3 model, reporting search revenue up 17% year-on-year, cloud revenue up 48%, and a backlog increasing 55% quarter-on-quarter to $240 billion, boasting the most solid distribution moat (Gemini monthly active users over 750 million, collaboration with Apple covering ~2.5 billion devices). Amazon Web Services (AWS) AI annualized revenue has surpassed $15 billion, with self-developed chips (Graviton/Trainium) annualized over $20 billion, possessing the strongest cost control potential in the wake of the DeepSeek effect. Microsoft's Copilot paid share dropped from 18.8% to 11.5% within six months, indicating short-term commercialization lag, but its "dual-layer monetization" structure (M365 subscription + Azure compute) retains medium to long-term value. Overweighting stocks with clear distribution capabilities and verifiable CapEx-revenue correlation is suggested, while avoiding pure infrastructure investment companies.
**Adtech: The Programmatic Advertising Market is Undergoing a Profound "Squeezing of Excess," with Budgets Migrating from Open Web to In-App Closed-Loop Platforms; Platforms with First-Party Data Flywheels and Algorithmic Moats Will Continue to Benefit.** The Open Web programmatic advertising market is in structural decline (annual growth ~3%, significantly lower than the 10%+ for Walled Gardens), rooted in principal-agent misalignment—for every dollar an advertiser spends, only about $0.45 reaches the publisher. Although traffic from MFA (Made for Advertising) websites has declined from a peak of 30%-40% to 15%-20%, it still erodes advertising efficiency. In the post-cookie era, alternative solutions like UID 2.0 have limited reach and face challenges in multi-party interest coordination, exacerbating signal loss issues in the Open Web. The core beneficiaries of budget migration are closed-loop platforms possessing first-party behavioral/transactional data. AppLovin has established an algorithmic moat with its AXON 2.0 AI engine, reporting full-year 2025 revenue of $5.48 billion (year-on-year +70%) and an adjusted EBITDA margin as high as 82%. Its e-commerce self-service platform is in limited beta testing, with only about 1.3% of available ad impressions monetized through the AXON system, indicating vast penetration potential, though risks related to an SEC compliance investigation warrant attention. After a strategic restructuring, Unity's Vector AI advertising engine has achieved a qualitative leap, accounting for nearly 80% of its Strategic Grow revenue and posting mid-teens quarter-on-quarter growth for three consecutive quarters. The upcoming Q2 Runtime data integration is expected to initiate a "compound interest flywheel," establishing a turnaround narrative, though with lower conviction than AppLovin. Looking forward, the marginal utility of data assets is increasing against the backdrop of tightening privacy regulations. Data trading opportunities between Walled Gardens will create new value distribution patterns. Focusing on leading Adtech stocks with deep algorithmic moats and clear commercialization paths is recommended.
**Risk Warning: Business development may fall short of expectations; Industry growth may fall short of expectations; Regulatory uncertainty; Technical risks; Commercial deployment risks; Other risks.**
Comments