lulusarcade| Hong Kong Stock Connect is on vacation. Who is violently boosting Hong Kong stocks?

　　炒股就看金麒麟分析师研报lulusarcade，权威，专业，及时，全面，助您挖掘潜力主题机会！

　　港股的基金重仓股的活跃度此前高度依赖于A股基金、沪港深基金，并在过去很长一段时间出现若港股通因假期休市则港股表现难堪的局面，而今在缺乏A股基金、沪港深基金的资金的五一假期，港股不依赖于A股基金的大幅拉升或意味着外资机构和QDII基金正在大力回补仓位，五一假期之前多个港股公司均出现机构、境外主权基金增持股份的现象。有基金经理强调五一期间的服务业、零售业数据或成为外资观察中国宏观经济的一个角度。

　　滞后一年，港股AI大模型板块起飞了

　　受境外资金挖掘洼地中的港股AI大模型板块，以及外资观察中国五一期间经济复苏情况的乐观判断，香港$恒生科技指数 （800700lulusarcade.HK）$飙升超4%，带动AI核心品种全线起飞。

　　尽管AI大模型在A股和美股的炒作时间已持续一年，但港股的AI大模型板块却在过去一年间成为许多公募基金经理的阴影，券商中国记者注意到，在A股和美股AI大模型热火朝天的炒作时期，不少错过行情的基金经理南下港股挖掘“更便宜、更正宗的人工智能”，但与基金经理预期相反的是，港股市场不仅对AI大模型在全球的火热无动于衷，更反向出现深度暴跌的罕见景象，这导致许多南下挖掘“正宗标的”的公募在2023年以及2024一季度期间备受打击。

　　而今，当A股和美股的AI大模型炒作告一段落之际，包括A股基金、QDII基金以及境外主权基金等全球资本开始加速流入香港市场尤其是科技板块。5月2日，前期被公募基金大量减持和低配的港股上市的科技股全线拉升，显现出机构资金回补的逻辑。

　　券商中国记者注意到，公募前期减持越多，当下回补资金越猛，截至5月2日收盘，去年四季度被基金减持干净的商汤科技当日大涨超36%，8个交易日内累计涨幅高达1.7倍。在此前的连续杀跌行情中，从十余只公募基金重仓到如今剩下一只基金独苗的$微盟集团 （02013.HK）$也大涨14.39%。根据基金一季度报告，目前仅剩下嘉实港股互联网基金持仓该股。

　　此外，公募基金挖掘的AI概念股$金蝶国际 （00268.HK）$5月2日涨超11%，在AI领域与商汤科技、$优必选 （09880.HK）$齐名的$第四范式 （06682.HK）$大涨超14%，与华为、腾讯分别合作医疗大模型和人工智能实验室的$医渡科技 （02158.HK）$涨近5%，8个交易日已为基金经理带来30%的收益，发布AI大模型的$美图公司 （01357.HK）$在多只明星基金的资金回补下亦上涨约4%，在新质生产力、工业机器人和AI大模型三大概念刺激下，针对制造业赛道发布生成式AI工业制造大模型的$创新奇智 （02121.HK）$，亦随着机构资金回补而飙升约15%；而在2023年年度报告中宣布医疗AI大模型和生成式医学AI已首次实现商业化落地的$医脉通 （02192.HK）$，此前亦遭资金抛售，5月2日该股涨近5%，最新基金一季度报告显示，医脉通在AI大模型的加持下，今年一季度期间迎来国泰基金旗下公募产品的重仓。

　　A股基金假期暂别港股通，谁在补仓港股？

　　在过去一年间，每一轮假期因港股通暂停导致A股基金无法南下，导致港股在缺乏公募港股通资金的助力下，往往在假期成为杀跌对象，而今港股通暂停中的港股如此凶猛上涨的做多力量是谁？

　　“我们判断当前港股通暂停时的港股行情主要是外资机构和公募QDII资金在持续做多。”深圳一位通过港股通重仓港股的基金经理认为，此前已有大量外资机构在今年一季度开始买入观察仓，同时QDII基金的港股仓位在过去三个月的时间内一直在稳步增加，但由于港股仓位在许多公募QDII仓位占比中仍然较低，这也意味着这种增仓的空间还比较大。券商中国记者也注意到，从600亿港元市值跌至50亿港元市值的AI医疗大模型公司医渡科技，4月24日发布的港股权益披露信息显示，国际主权基金BIA公司增仓该股至5360万股。而微盟集团在5月2日的大幅拉升亦因4月30日晚间所发布的公告，腾讯公司以定增方式增持这家向AI赛道转型的技术营销公司，并使得微盟集团补充约9亿元的资金。

　　显而易见，五一期间A股基金无法通过港股通南下的背景下，港股的强力上涨更隐含出当前港股市场已不仅仅是内地公募机构看好趋势行情，海外资金的回流亦与内地机构保持一致。港股华宝港股互联网基金经理丰晨成认为，从中国经济复苏节奏上看，宏观面的预期仍是分子端扰动投资者情绪的因素，五一期间的服务业、零售业数据或成为外资观察中国宏观经济的一个角度。

　　业内人士亦强调，5月2日港股通暂停中的科技指数暴涨局面，反映出当下港股的风险偏好开始大幅增加，并已从内地公募单方面看好市场转向A股基金、港股基金、QDII基金和外资机构的一致性做多氛围。

　　关于港股市场尤其是连续下跌时间一年甚至三年的科技股，公募基金究竟如何看待姗姗来迟的港股AI行情？

In this regard, Southern Hong Kong growth Fund Manager Wang Shicong believes that the global AI cycle is still in the early stage, and AI will become a key factor in improving social production efficiency in the next 10 years. In terms of hardware, GPU is still the main line, and the demand side is still on the demand side. Not only GPT, Claude, Sora, Gemini and other big models are still being updated rapidly, but there are also more applications on the reasoning side. $NVDA.US $released the new B-Series GPU at the GTC conference, which has greatly improved its efficiency and architecture. It is especially suitable for the application of the inference side, and brings great opportunities to the whole upstream industry chain. Most of Nvidia's upstream industrial chains are in mainland China and Taiwan. In addition to servers, more end-to-side computing power will begin to gradually follow the rise of AI mobile phones and AIP C, and may further accelerate from the second half of 2024 to 2025, which will lead to a new replacement cycle in the next 2-3 years. At the same time, this will also bring investment opportunities for advanced manufacturing. It is expected that AI technology will be gradually applied to all kinds of vertical applications. Graphics and text, video, autopilot, recommendation algorithms, education, databases and other fields will usher in breakthroughs in 2024.

Liu Jie, manager of Ping an Shanghai, Hong Kong and Shenzhen Fund, believes that the main factors of the trend of the equity market in 2024 include the domestic and foreign economic situation, the international monetary environment, international trade relations, and so on. The medium-term trend of stabilizing and improving the domestic economy remains unchanged, while external demand is currently bullish than market expectations, providing more opportunities for China's manufacturing industry to "go out to sea" to find a new growth direction. the fund is also working hard to explore the subdivisions of China's manufacturing industry that are globally competitive and capable of global growth. Once the international monetary environment starts the easing cycle of major countries, the valuation of the global capital market is expected to be marginal repaired, and the fundamentals will also improve marginally; under the background of this year's global "election year", there may be certain disturbing factors. It is believed that under the general trend of international division of labor and globalization, the international market will be an important growth and development direction of China's manufacturing industry. The equity market may still show certain differentiation characteristics in 2024. On the one hand, the market will continue to look for relatively certain cash return opportunities in low-valuation and high-dividend varieties with defensive attributes, such as public utilities, big finance, transportation, and resource stocks; at the same time, the long-term innovative growth direction represented by the digital economy, intelligent driving and robots will gradually emerge after the relevant varieties' valuation has been fully adjusted.

Deng Xinyi, general manager of the research department of Noan Foundation, also said that the demand side of AI artificial intelligence, all kinds of applications emerge one after another, especially the recent domestic large models represented by KIMI continue to iterate, and see the dawn of the development of domestic artificial intelligence. I believe that with the continuous improvement of the global large model ability, the office, programming, marketing and other scenarios of To B side. To C-side autopilot, picture and text generation, games and other areas are expected to usher in a rapid increase in penetration, AI PC / mobile phone / MR and other hardware innovations are expected to further promote AI to enter the positive feedback of the iteration of capabilities. In addition, at the policy level, the government work report puts forward the "artificial intelligence +" action, and the relevant policies and strategic layout also clearly put forward the requirements to accelerate the development of artificial intelligence. Therefore, with the continuous improvement of the ability of large models, it is expected to usher in the explosive growth of computing power and application demand.