A：In 2026, China’s scientific and technological development has reached a critical new starting point. We use the word “burning” to encapsulate today’s tech investment opportunities, which encompasses three dimensions: First, core technology sectors are poised for a cascading wave of breakthroughs. For instance, AI is set to profoundly reshape business processes; domestically produced semiconductors are advancing through自主创新; and commercial spaceflight is also approaching a critical inflection point, with multiple fields reaching the eve of explosive growth simultaneously. Second, the innovation ecosystem will shift from isolated breakthroughs to a sweeping, chain-reaction momentum driven by full-value-chain collaboration. China’s unique fuel infrastructure, abundant engineering talent pool, extensive end-to-end industrial chain, and vast ultra-large market will collectively propel comprehensive breakthroughs and tangible outcomes in technological innovation. Finally, guided by the 15th Five-Year Plan, China’s tech sector carries both the nation’s aspirations for the future and a strong sense of responsibility. Breakthroughs in innovative pharmaceutical companies, as well as the establishment of new standards in humanoid robotics and brain-computer interfaces, exemplify how our expertise aligns with this blazing torch of technological innovation.

A：We firmly believe that AI is not just a tool but also an accelerator of technological research and development. It can significantly enhance efficiency, reduce costs, shorten R&D cycles, and mitigate uncertainty, providing comprehensive support for technological innovation. This catalyzes explosive breakthroughs in technology, unlocks growth potential across the tech industry, and creates long-term, highly predictable investment opportunities for capital markets. In response to shifting macroeconomic conditions and the deepening alignment of investment themes with industry trends, Noah Asset Management remains committed to in-depth research. By hosting this strategy conference, the firm aims to decode the underlying logic behind China’s tech boom and identify high-potential investment opportunities. Through macroeconomic analysis, report reviews, and roundtable discussions, Noah seeks to help investors navigate industry volatility, distinguish true value, and seize certainty-driven opportunities.

A：Ms. Yi Huan believes that the global economic recovery and supply shocks are the key issues this year. Despite market attention being drawn to events such as the U.S.-Iran conflict, China’s economy has exhibited signs of cyclical recovery since the start of the year. In particular, the real estate cycle adjustment is nearing its end, while efforts to cultivate new growth drivers are beginning to lay the groundwork for a sustained rebound. Furthermore, China’s resilience to external shocks has strengthened compared with the period of the Russia-Ukraine war in 2022. Meanwhile, commodity and bond pricing is path-dependent, whereas equity assets place greater emphasis on the convergence of probability distributions. As the market impact of the U.S.-Iran conflict becomes clearer, equity markets are likely to revert to some of the key themes that prevailed at the start of the year and undergo reassessment and rebalancing.

A：The two axes refer to the two coordinate axes represented by a 45-degree angle: one axis shows the growth rate of fixed assets, and the other shows the growth rate of actual operating revenue. If the data points lie below the 45-degree line, it indicates that capacity expansion is outpacing demand growth; conversely, if they lie above the line, demand expansion is outpacing capacity expansion.

A：From 2021 to before 2024, most sub-sectors were undergoing capacity reduction or contraction and were unable to undertake capital expenditures. However, after the third quarter of 2024, demand growth in half of the industries began to outpace capacity expansion, signaling that China’s economic cycle has entered a moderate recovery phase this year, with the manufacturing sector in particular emerging from deflationary conditions.

A：At present, it is inadvisable to make major directional decisions regarding the real estate market. Particularly in the absence of large-scale demand-side policy stimulus, the property cycle is unlikely to stage a V-shaped recovery following its sharp correction. However, compared with previous years, the share of real estate construction investment in GDP has already declined significantly. Even if it continues to fall, its impact on the overall macroeconomy will remain relatively limited. Meanwhile, many sectors have already emerged from the deflationary spiral at the aggregate level. The real estate sector’s adjustment has gone even deeper than the trough experienced during Japan’s Lost Two Decades, and its share of GDP in investment is now close to the U.S. low point. As a result, real estate no longer constitutes a drag on the overall economy.

A：We had originally projected that the PPI would turn positive in May or June. However, it now appears that the turnaround will occur as early as March, primarily driven by demand outpacing supply growth and a restoration of pricing power, rather than imported inflation or rising oil prices. This upgrade carries tremendous significance for the revaluation of China’s economic assets, particularly in the manufacturing sector, and is reflected in changes across multiple dimensions, including corporate profits and the RMB exchange rate.

A：In the past, when China’s GDP targets were not particularly narrow, there was a significant divergence between real GDP and nominal GDP. A 0.5-percentage-point change in real GDP could translate into a 2 to 3 percentage-point increase in nominal GDP. Moreover, corporate profits exhibited roughly twice the elasticity relative to nominal GDP, meaning profit growth could reach 5 to 6 percentage points. Furthermore, the RMB exchange rate has also exhibited pronounced procyclicality. Together, these factors mean that economic forecasts for China tend to be closely tied to cyclical fluctuations.

A：The RMB exchange rate is currently at a medium-to-long-term inflection point, and is expected to reach around 6.6 by year-end. As China’s private-sector GDP growth picks up and the renminbi appreciates, Chinese companies’ dollar-denominated earnings are poised for a substantial boost, presenting a rare and highly certain growth opportunity on a global scale. Meanwhile, with the global capital expenditure cycle gaining momentum—driven in particular by rising demand in the defense industry and reindustrialization efforts—China’s nominal GDP growth is expected to outpace that of the United States for several consecutive years.

A：Currently, supply-side challenges are primarily attributable to protracted development cycles and supply-side constraints in China’s real estate sector, particularly with respect to consumable materials. Meanwhile, infrastructure such as the power grid in developed countries worldwide, including the United States, has not experienced significant expansion over the past three decades. Most consumable materials are essentially being depleted through depreciation. In China, demand for construction-related inputs peaked during the property boom and has since tapered off as the sector undergoes de-real-estateization.

A：Steel prices have fallen from a high of 40 points to the current level of 14 points, primarily due to shifts in demand related to the construction and real estate sectors. Steel prices have already fallen to their lowest possible level. While the decline in copper demand has been less pronounced than that of steel, it too is being affected by China’s property-related demand.

A：When global consumable steel is categorized into China‑real‑estate‑related and non‑related segments, the share of China‑real‑estate‑related steel consumption in global total has declined from a certain level to 7%, indicating significant room for further growth. However, if Chinese real estate demand is excluded, global consumable consumption would still have grown by 2 to 3 percentage points annually between 2020 and 2024.

A：Although global fiscal policy remains accommodative, it is gradually encountering a tightening constraint: as monetary conditions continue to ease, a growing number of market participants are becoming reluctant to hold bonds as safe-haven assets. Bond yields have underperformed hard assets, leading to a surge in fiscal leverage to its highest level in 200 years, which will soon feed back into higher bond yields.

A：Currently, three major interwoven cycles are at play: fiscal unsustainability leading to currency depreciation, the restructuring of the global order (reindustrialization of the defense industry and supply chain security), and the AI capital cycle (generalization and accelerated implementation of applications). This year presents a favorable opportunity for asset allocation, and the outbreak of a U.S.-Iran war could further reinforce these trends.

A：In the short term, the closure of the Hormuz Strait is likely to have only a limited impact on the global economy, as the shipping system is resilient and many vessels had already departed in advance. However, if the supply gap persists for more than one year, a cliff-like plunge in demand similar to that seen during the Iran–Iraq War could occur, with severe repercussions for the global economy.

A：Depending on the circumstances, the forecasts may be adjusted. In particular, losses in Asian countries could increase, while the financial stability of Gulf states may face prolonged challenges. Across different quadrants, sectors with abundant resources and those experiencing rapid improvements in productivity will be the primary focus.

A：When cash flows in the Gulf states are under pressure, they may tactically reduce their demand for gold or even engage in selling, which could lead to a correction in gold prices. However, in the longer term, gold demand may strengthen, as a higher oil price floor will boost consumer spending. That said, for both China and the United States, nominal GDP growth could offset some of this impact, preventing a sharp decline in consumption.

A：Last year, liquidity conditions were exceptionally favorable, with both the U.S. dollar index and oil prices declining. This drove global risk assets to outperform expectations. This year, however, the U.S. dollar index has risen. Although oil prices have shifted up slightly, overall liquidity is weaker than last year, leading to greater market divergence. The divergence between sectors driven by fundamental strength and those propelled by liquidity will become even more pronounced.

A：This year, the rise in the U.S. dollar index and oil prices will partially offset the liquidity easing driven by fiscal stimulus, thereby intensifying divergence across countries, industries, and sub-sectors with varying cash-generating capabilities. Moreover, rising oil prices will increase the fiscal burden of government subsidies, thereby partially offsetting the stimulative effects of fiscal easing.

A：Three major structural trends may be strengthening: First, following the U.S.-Iran conflict, fiscal easing has intensified globally, driven by increased defense and military-industrial spending and post-disaster reconstruction needs. Second, as the global order is being reshaped, there is a renewed push for reindustrialization in the defense sector and heightened emphasis on national security considerations. Third, although AI-related capital expenditures are facing marginal weakening due to rising costs and reduced cash flows in Gulf states, the practical implementation of AI applications is accelerating, propelling the arms race into a self-accelerating phase at a higher technological level.

A：In the short term, China has demonstrated strong resilience to this energy diversification shock, primarily due to its robust preparedness and relatively low dependence on oil. However, if the shock persists for an extended period, a contraction in global demand will adversely affect China’s terms of trade and certain profit-sharing arrangements. In the long term, China will benefit from the accelerated adoption of new energy technologies and heightened global awareness of energy security. In particular, new energy-related sectors such as energy storage are expected to experience faster growth, enabling China to increase its share in global manufacturing.

A：When confronted with supply-side shocks, China can effectively respond by deploying supply-side policies, such as accelerating the energy transition, reducing energy intensity, and expanding new energy capacity. This shock will strengthen China’s investments in high-tech industries, national security, and the green energy transition, and is expected to further increase Chinese manufacturing’s global market share by leveraging its advantages in cost control, supply-chain resilience, and operational efficiency.

A：Test-time scaling e is a reasoning-based scaling technique that increases computational effort during inference to improve model performance, enabling model fine-tuning without relying on complex technical patches.

A：The main challenges include a lack of high-quality private-domain data with diverse causal logic, as well as a fundamental shift in ROI logic. Enterprises are increasingly focused on whether AI can generate tangible incremental profits, which necessitates a higher level of model capability.

A：Currently, AI has entered a new stage where professional services can be delivered on a large scale and systematically, thereby transforming into productivity. For example, in the AI for business field, AI has significantly lowered the threshold for professional services in areas such as programming, law, education, and marketing, and has demonstrated efficiency improvements through examples like GitHub Pilot and PwC’s legal review services. AI is transforming R&D from a traditional trial-and-error process based on experience into a computable and predictable one. For example, Google DeepMind’s crystal structure prediction model has significantly improved research efficiency, signaling that in the future, AI will become an R&D entity capable of independently formulating hypotheses, conducting validations, and reaching conclusions.

A：AI assistants have shifted from passively following instructions to coordinating tasks across multiple scenarios, and from merely transmitting content to acting as service executors. Particularly in the field of autonomous driving, there is a transition from rule-based systems to data-driven ones, which has enhanced generalization capabilities in complex environments and boosted productivity in the transportation industry.

A：The key turning point in edge-side development lies in hardware reconstruction. A wide variety of AI-enabled devices have emerged, including AI phones, IPCs, and AI glasses. Among these, the share of AI chip shipments has been steadily increasing. AI hardware is now deeply integrated into various aspects of daily life, such as emotional companionship and health management.

A：The core driver of AI infrastructure lies in innovation in underlying hardware. NVIDIA leads in the GPU sector, while CPU manufacturers are developing their own dedicated computing chips to reduce costs and achieve better compatibility. The overall industry trend shows that ASIC chips and GPU chips coexist and develop together, while the demand for CPUs has increased substantially as AI applications deepen. Furthermore, network interconnection technologies are evolving toward larger-scale supernodes and cross-node connectivity to enable flexible and efficient computing resource scheduling.

A：There are two main technical approaches in the industry for addressing the memory wall problem. One approach is for NVIDIA to build a land cache pool, placing the KV key-value cache next to the computing chip. This significantly alleviates the workload by storing contextual data in an AI-native infrastructure architecture, thereby enabling efficient data processing at the chip level. Another technical approach is the CSB cloud platform, which adopts a C‑fork L‑memory‑pool architecture to enable platform‑based operations. It dynamically allocates DRAM and other storage resources based on demand, thereby meeting the storage requirements of diverse use cases.

A：The current data complexity is reflected in the continuous development of AI systems, ranging from the original large-model infrastructure for training to potential future directions—such as integrated storage, computing, and inference, integrated storage and computing, or decoupled storage and computing, among other diversified approaches. Data layering is prominent: different chips and memory dies have distinct roles and characteristics. In the future, various memory chip dies may be strategically placed around the main chip based on their specific attributes, enabling collaborative operation with SRAM, 3D DRAM HBM, and 3D stacked HBF, thereby maximizing the energy efficiency of GPU-specific accelerator chips.

A：Currently, domestic foundries have achieved stable mass production on more advanced process nodes and have entered the scale-up phase for the next-generation advanced processes. China has largely completed the domestic substitution of its entire semiconductor industry chain, spanning design, manufacturing, packaging, equipment, and materials. In certain areas, such as photoresists, it has achieved a breakthrough from having none to developing viable products. It is projected that by 2027, China’s chip self-sufficiency rate will increase significantly to around 30%, while both the size and market share of the domestic semiconductor equipment and materials markets will expand substantially. Meanwhile, at the national level, industrial capital, such as the third phase of the Big Fund, is making targeted investments, further clarifying the direction of domestic semiconductor investment. In particular, in the areas of AI chip manufacturing and expansion of memory chip production capacity, China is advancing on a path of independent innovation and self-reliance.

A：Globally, semiconductor sales have seen a sharp increase driven by AI. Since 2020, the United States has tightened its restrictions on China’s semiconductor industry. However, these measures have also prompted China to establish a robust support system encompassing policy frameworks, financial resources, and talent development. The establishment of the STAR Market has addressed the financing challenges faced by the semiconductor industry, which typically require substantial capital investment but offer long payback periods. As a result, numerous semiconductor companies have listed on the STAR Market and secured significant funding, thereby driving substantial advances in domestically developed, self-reliant technologies.

A：By 2025, China’s space program aims to transform rocket manufacturing from a handcrafted, artisanal process into a mass-production, assembly-line operation through technological breakthroughs. Whoever succeeds in achieving both rocket reusability and scalable production will hold the key—a veritable ticket to the trillion-dollar orbital economy.

A：As transportation capacity ceases to be a constraint, the ceiling for commercial spaceflight will be lifted. Satellite communications will be dramatically enhanced, with direct satellite connectivity enabling mobile phones to eliminate signal blind spots worldwide; suborbital transport capable of reaching any point on Earth within an hour will reshape the logistics landscape; and space-based pharmaceutical manufacturing and in-orbit factories will leverage the microgravity environment to produce new materials that cannot be created on Earth. These once-futuristic concepts are now transitioning from science fiction into reality.

A：In 2026, China’s space program will enter a phase of comprehensive breakthroughs, while the humanoid robotics industry will undergo fundamental transformation. ChatGPT is driving AI development, and humanoid robots are giving AI a physical form, ushering it into mass production. Taking Tesla’s Optimus robot as an example, if its cost can be reduced below US$20,000, it could unlock a market worth trillions of dollars, fundamentally reshape manufacturing cost structures, and become the centerpiece of the Fourth Industrial Revolution.

A：The key to technological breakthroughs lies in the “hand”—specifically, the development of dexterous robotic hands. The evolution from simple grippers to multi-degree-of-freedom, human-muscle-inspired actuators represents not only a change in mechanical structure but also a breakthrough in algorithms and control capabilities. Dextrous hands serve as the ultimate interface for AI to connect with the physical world. Mastering dexterous hand technology means robots can move beyond merely imitating humans and achieve superior performance.

A：In addition to dexterous hands, rapid advancements in technologies such as electronic skin and sensors are essential to equip robots with perceptual capabilities. Meanwhile, robotic materials are also evolving, enabling robots to achieve enhanced functionality and lighter, more agile structures. Technological breakthroughs in these niche areas will drive the next stage of robotic evolution.

A：The past decade has been a period of profound transformation for China’s innovative pharmaceutical industry. Policy reforms, ecosystem development, and changes to the pricing mechanism have collectively driven substantial growth in the sector. In 2025, innovative pharmaceutical companies are shifting from reliance on capital infusions to financing based on their proprietary technologies and product pipelines, signaling a transition in the industry lifecycle and a repositioning of valuation benchmarks. Over the next few years, the market will place greater emphasis on companies’ ability to monetize their assets and on the enhancement of their pricing power for global innovative pharmaceutical assets.

A：In 2021, a breakthrough ADC drug was unveiled at an academic conference, presenting data that are poised to reshape the diagnosis and treatment landscape of breast cancer. Currently, China has seen significant improvements in the ADC field, with substantial increases in the number of R&D pipelines, clinical trials, and global market share. Moreover, breakthroughs have been made in the development of novel molecular entities such as dual-target and dual-cargo ADCs, highlighting China’s strengths in innovative drug innovation.

A：Similar phenomena may emerge. In the future, secondary-market investors could view the innovative-drug sector as having entered a “Zhang Xue locomotive moment.” There might even be a proprietary term coined to describe this distinct phase in China’s innovative-pharmaceutical industry, driven by the country’s comparative advantages in engineering capabilities, an abundant pool of engineering talent, robust clinical resources, and a bold, go‑for‑it entrepreneurial spirit.

A：Doing well in the domestic market is key to successful international expansion; you can’t hammer iron unless your own tools are strong. Only by strengthening the domestic market can a solid foundation be laid for global expansion. Moreover, given the substantial room for growth in the penetration of innovative drugs in China, the development of the domestic market is of paramount importance.

A：The innovative pharmaceutical industry has now entered a phase of high-quality, rapid development. Despite volatility in the secondary market, it is essential to adhere to the sector’s inherent characteristics: high capital intensity, high risk, high potential returns, and extended development timelines. At the same time, it is important to acknowledge the industry’s current stage of development and shortcomings in basic research. Greater attention should be given to ensuring fair trading practices and fostering a more mature investment mindset. Over the next few years, the potential for returns will depend on the successful achievement of key milestones and progress in global commercialization.

A：Brain-computer interfaces are currently still in the thematic investment phase. However, they exert strong externalities on supporting industries, and certain niche segments may begin to deliver results that align with investment frameworks ahead of others. When selecting a brain-computer interface company, focus on those with a clear niche, a breakthrough differentiator, and the ability to set industry standards. The brain-computer interface ecosystem may be approaching a tipping point. In the early stages, applications will primarily focus on serious medical use. The next three to five years will be characterized by intense R&D efforts and a race to accumulate clinical evidence. After five years, consumer-oriented applications are expected to begin gaining traction. Over the longer term, BCI is likely to mutually reinforce embodied AI, together driving industry growth.

A：In the global energy system, fossil fuels still dominate; however, oil and natural gas reserves are expected to be depleted within a century, while coal reserves are projected to last for approximately 200 years. Meanwhile, to achieve the goals of the Paris Agreement, the global community must limit temperature rise and advance carbon neutrality, which has resulted in carbon taxation systems that impose institutional constraints on fossil fuel use. Variable power output from wind and solar generation poses challenges to the power grid, leading to a sharp deceleration in their deployment. As a result, the renewable energy sector now faces new hurdles.

A：Nuclear fission faces a dual challenge: the finite nature of uranium resources and the management of spent nuclear fuel, making it difficult to serve as a primary energy source for large-scale deployment. Globally proven, economically recoverable uranium resources are estimated to last approximately 100 years. Without an increase in the number of operational fusion reactors, continued expansion of nuclear power would accelerate the depletion of these resources. This is because the most common energy source in the universe is controlled nuclear fusion, which boasts virtually unlimited fuel reserves, intrinsic safety, and the advantages of being clean and carbon-free. It is often referred to as the “artificial sun” and the ultimate energy solution.

A：The development of controlled nuclear fusion will give rise to a new industrial value chain, spanning upstream core materials—such as heavy water, high- and low-temperature superconducting magnets, and specialized alloys—through midstream critical equipment—including vacuum vessels, divertors, fuel-handling systems, and steam generators—to downstream power-generation applications.

A：The core principle of nuclear fusion is the combining of two hydrogen atoms to form a heavier atomic nucleus, releasing energy in the process. For example, the deuterium–tritium fusion reaction produces helium, which is environmentally benign, and it boasts a high energy density. Moreover, Earth’s seawater contains abundant supplies of these fusion fuels.

A：China holds a world-leading position in global research on controlled nuclear fusion. On Hefei Science Island, a unique global R&D cluster has emerged, bringing together world-class facilities such as EAST, BEST, and COAST. Meanwhile, large-scale experimental devices are under construction in Shanghai, Chengdu, Nanchang, and other cities, creating a robust competitive landscape characterized by state-led leadership and active participation from private enterprises.

A：China’s AI sector presents exceptionally strong investment opportunities over the next 1-2 years, primarily in three areas: first, a boom in AI application industries, encompassing both the reshaping of consumer-facing traffic models and the exponential growth of AI-native enterprise applications; second, a comprehensive upgrade of the semiconductor industry in advanced manufacturing processes driven by technological breakthroughs and surging AI application demand; and third, a significant positive surprise—China’s tech sector, particularly AI applications, is expected to become more competitive on the global stage, with robust performance in overseas markets. Meanwhile, the emergence of open cloud platforms will accelerate the globalization of Chinese tech companies and lead to a reevaluation of their valuations.

A：AI is currently regarded as the core driving force behind this technological revolution, a view on which tech giants in both China and the United States largely agree. Despite substantial capital expenditures and the current lack of widely recognized hit applications, a long-term historical analysis reveals that AI is a technological revolution we cannot afford to lose. The positive correlation between computational power investment and large-model AI capabilities is evident, and it holds the potential to achieve artificial general intelligence (AGI), thereby securing global leadership in technology. Therefore, AI investment is highly tempting and constitutes an arms race.

A：Looking ahead to 2026, areas such as autonomous driving, robotics, and AI agents are highly worth paying attention to. With the advancement of large-model technologies, AI capabilities are gradually maturing. What is now needed is robust engineering expertise to closely integrate AI with a wide range of application scenarios. Moreover, 2026 will be a pivotal milestone for the full realization of performance gains across NVIDIA’s supply chain. At the same time, domestic Chinese manufacturers in the computing-power segment are expected to see a substantial increase in output as bottlenecks in advanced process technology and packaging are resolved.

A：The growth logic of AI is not the traditional software application model, but mainly new application forms such as models and products, and agent products. The spiraling, symbiotic relationship between AI and hardware is a technological shift with a high degree of certainty. From the PC internet era to the mobile internet era and now to AI, the development of hardware and software has been deeply coupled, mutually defining each other, and jointly determining the boundaries and ceiling of information technology capabilities.

A：AI has gone through five stages: software exploration (represented by Google’s Transformer architecture, which provided a theoretical foundation for handling massive datasets); the explosion of groundbreaking software products (such as the ChatGPT conversational application based on GPT-3.5, demonstrating the emergent capabilities of large models); hardware bottlenecks (NVIDIA launched dedicated GPU chips tailored to the Transformer architecture, significantly boosting computing power); hardware adaptation and performance enhancement (chip companies have upgraded computing power and bandwidth to meet new algorithmic demands, leading to the emergence of products like LPUs and high-speed optical modules); and software adaptation and further proliferation (as foundational large models continue to emerge, they are placing new requirements on chips, spurring even more innovative products and services).

A：Yes, AI is undoubtedly a real productivity revolution, which I liken to the amplification of mechanized intellectual labor. Within my investment framework, I focus on and allocate capital along the trajectories of continuous productivity enhancement and rapid technological iteration, with a primary emphasis on two areas: first, the development of large-scale AI models; and second, the infrastructure hardware stack that aligns with the computational demands of these models.

A：In the competitive landscape of large models, the global tier-one cohort of leading models has largely been solidified. The competitive barriers for these leading players primarily stem from two factors: first, the ability to sustain substantial investment; and second, access to proprietary data assets. Meanwhile, algorithms are continuously iterated to enhance model capabilities. As these capabilities improve, the interaction between humans and AI is expected to be reshaped, with the emergence of proactive agents being particularly promising.

A：Computing power is the foundation of both model training and inference applications. At the foundational model training stage, ubiquitous and general-purpose computing power is required to support diverse algorithmic experiments and innovations at the lower levels. Meanwhile, at the inference and application layers, as adoption rates continue to rise, the demand shifts toward cost reduction and capacity expansion, along with latency reduction, in order to enable large-scale deployment of AI applications.

A：China has four major advantages in the global AI wave: the richest AI application ecosystem, the world’s strongest engineer dividend, a complete hardware supply chain, and strong implementation capabilities. Despite the difficulties and challenges in the hard-tech sector, China’s leading large-model companies have consistently remained at the forefront of R&D as they compete with the world’s top models in terms of capabilities.

A：Semiconductors, as a core component of hard technology, serve as the foundational infrastructure of the digital world, encompassing a wide range of chip types, including computing chips, memory chips, analog chips, and power chips. Compared with cyclical industries, the semiconductor sector serves as the most critical infrastructure component in each innovation cycle. Its investment appeal lies in thoroughly understanding marginal shifts at the industry’s micro level, with a focus on the competitive dynamics and technological barriers across various segments—design, manufacturing, packaging, and supporting industries. To build a moat in semiconductor investing, one must seize on micro-level changes in the industry and adhere to a long-term investment philosophy, paying particular attention to the sequence and growth trajectory of the semiconductor industry’s comprehensive upgrade driven by AI. At the demand level, identify the core value-adding segments; at the supply level, analyze the supply–demand dynamics of both advanced and mature semiconductor manufacturing processes; at the industry-chain level, examine the shift from global division of labor to regional restructuring. Meanwhile, focus on key areas such as enhancing chip design capabilities, upgrading chip manufacturing capacity, and achieving breakthroughs in equipment and materials.

A：In hard-tech investing, we begin by adopting a cyclical perspective to analyze investment opportunities in technology, identifying blockbuster products that align with the current era. We then focus our efforts through a dual domestic–international positioning strategy. The specific approach involves using technology maturity curves (such as the Gartner Hype Cycle) to assess the stage of a technological breakthrough and its value, while also taking into account the feasibility of commercialization, economic viability, and barriers posed by the competitive landscape. Secondly, pay attention to the technology adoption lifecycle curve to observe the engagement of different audience segments and the development of the technological ecosystem at each stage. Finally, during the explosion phase that follows widespread product adoption, make investment decisions by referencing the product life-cycle curve.

A：By overlaying the technology breakthrough curve with the ecosystem development curve, and the technology adoption lifecycle curve with the product lifecycle curve, one can identify the overlapping segment where both are advancing rapidly and are likely to generate excess returns—the so-called “double-high” phase. This allows investors to capture both the premium associated with rising optionality and genuine operating earnings growth (the “Davis double-click”).

A：The principal contradiction in China today centers on “mental productivity.” Attention is focused on the strategic emerging industries highlighted in the Government Work Report, including five major sectors—new energy, new materials, biotechnology, artificial intelligence, and robotics—and their respective sub-sectors. Abroad, the principal contradiction centers on the competitive yet cooperative relationship between China and the United States. Industry sectors in the technology field are categorized into three groups: those where China currently leads and is expected to remain dominant; those where the U.S. currently holds the lead but will face intense future competition; and those where the U.S. currently leads and is likely to maintain its advantage going forward. Different investment strategies are formulated for each of these categories.

A：When selecting flagship products, the key is to analyze the competitive landscape and assess the nature of the competitive advantage as well as its expected duration. Flagship products are not merely vehicles for specific technologies or brand credibility; they also boast substantial growth potential, help mitigate the risk of misjudgment, and enable investors to pursue alpha returns by capturing beta-driven gains.

A：The core investment themes in hard technology include: 1) foundational technological innovation centered on semiconductors, which ensures industrial security and enables systemic innovation; 2) mega-hit products that align with the needs of the times, such as robotics, commercial spaceflight, and innovative drugs, with particular emphasis on the explosive growth driven by AI; 3) the reconstruction of a new ecosystem within the global supply chain that balances both security and efficiency. Chinese hard-tech companies are actively embracing the global market and stepping onto the international competitive stage.

A：As technology investors, amid market volatility, we choose to navigate the ups and downs to identify and capitalize on the overarching patterns of technological evolution and emerging industry trends. While bubbles in the tech sector may fuel short-term growth, in the long run we need to look beyond the hype and invest in genuine industry trends. This approach will help ensure that, even if the sector’s bubble bursts, our investments can still deliver returns over the medium to long term. We manage portfolio-wide volatility risk through position sizing. For early-stage companies, we typically take relatively small positions, even though these smaller stakes can still offer significant upside potential. Furthermore, we align our investment horizon with long-term industry cycles, typically holding each stock for four to five years or even longer. Selecting high-quality companies significantly enhances both individual performance and the overall portfolio. When selecting companies, we ensure they possess the highest level of technological expertise, favorable technological breakthroughs and competitive positioning, as well as a clear business model.

A：In technology industry investing, we primarily focus on two types of companies. The first category comprises technology‑disruptive companies that redevelop cutting‑edge technologies in China to reach a level on par with, and ultimately surpass, global leaders. For example, many of the semiconductor sector’s star performers over the past few years have emerged against this very industry backdrop. The second category is AI-native companies. These companies do not have mature technologies either domestically or internationally, so we need to demonstrate a spirit of innovation. When investing in companies of this type, we require a full business-cycle loop to be demonstrated. We prioritize B2B enterprises and place particular emphasis on large-account order visibility.

A：During the investment process, we periodically invest in companies that the market perceives as overvalued or highly inflated. This is because, in the early stages of the tech industry, such companies typically command high valuations due to substantial R&D spending and long payback periods. We believe that companies with seemingly high valuations often receive relatively limited attention in the early stages. By accurately assessing and aligning with investors’ consensus views, it is possible to generate excess returns. When selecting companies in this space, we adhere to the following principles: highest level of technological expertise; the most favorable competitive landscape following a technological breakthrough; and a clear, easily monetizable business model.

A：When market sentiment diverges from industry trends, the primary task is to carefully examine the industry trends, analyzing the underlying drivers, the potential market size, barriers to entry in production processes, and the company’s competitive position within the industry. At the same time, assess the company’s core competencies, management and governance capabilities, as well as the alignment between its near-term performance and valuation over the next one to two years. Only through such thorough analysis and judgment can we overcome the negative impact of short-term market sentiment and improve the odds of making successful investment decisions.

A：At its core, the A- share market requires consideration of three key factors: industry trends, fundamentals, and market sentiment. Industry trends embody the tides of the times. In each bull market cycle, it is crucial to clearly identify and capitalize on the prevailing zeitgeist, aligning closely with the overarching direction of industrial development. Although fundamentals may lag industry trends, as the sector’s position becomes firmly established, the financial performance of leading companies will gradually come to the fore and remain stable over time. Therefore, during market volatility, it is crucial to identify the prevailing industry trends, re-engage with the emerging momentum, and make dynamic adjustments based on fundamental analysis. At the same time, while monitoring market sentiment as a reference point, one should avoid making it the central driver of trading decisions.

A：In tech investing, I deconstruct a company’s fundamentals by assessing its engineering execution capabilities and its ecosystem position within the industry. While current-period performance is important, the high capital expenditures typical of new-industry companies can lead to significant fundamental volatility. Therefore, I place greater emphasis on conducting in-depth industry-chain analysis to understand the root causes of such fluctuations in a company’s fundamentals. If the volatility is driven by macro factors or an increase in R&D spending, I wouldn’t be overly concerned. However, should the competitive landscape intensify or we find ourselves on the wrong technological trajectory, I would consider reducing my position.

A：As a fund manager, it is difficult to predict market sentiment, but it can be used as a trading reference. For example, when market sentiment is weak, the valuations and stock prices of high-quality companies may be undervalued. If your assessment of industry trends and fundamental factors is sound, this can be an opportune time to buy. Conversely, when market sentiment is overly bullish, one should be wary of risks. Overall, market sentiment should not be used as a core indicator for investment decision-making. To me, a company’s fundamentals and industry trends are more akin to the natural sciences: through rigorous analysis, one can arrive at answers that closely approximate objective reality. In contrast, market sentiment resembles the humanities—it is inherently subjective and even artistic—and it can significantly influence an investor’s emotional resilience and decision-making. When it comes to the sentiment of retail investors, particularly those acquired through digital channels, I tend to share more of my market insights to help them better understand and manage emotional swings, align their investment horizons, and make joint decisions about whether to continue our partnership.

A：Market sentiment is not merely noise; it may reflect underlying market dynamics, such as bull-bear cycles. Analyzing market sentiment can help identify these patterns and leverage their trends to inform investment decisions. Meanwhile, market sentiment can also serve as a trigger for entering or exiting trades, or for adjusting position sizes, and should be applied flexibly across different market phases.

A：My primary criterion is industry trends. If the industry’s growth ceiling is rising, I’m willing to accept short-term fluctuations in fundamentals. In fundamental analysis, it is important to distinguish between permanent losses and temporary transitional pains. Market sentiment should be used as a trading reference rather than the core basis for decision-making.