At the Future Industry Promotion Conference of the 28th Beijing International High-Tech Expo held on May 8, 2026, Gan Fei, General Manager of the Technology and Innovation Finance Department at HSBC China, delivered a speech. The following is a summary of her key points.

Gan Fei began by noting that when international investors look toward the future of innovation today, their questions have fundamentally changed from a few years ago. They are no longer just asking where growth is happening; they are more concerned with where talent is concentrating, where technologies are being commercialized, and which ecosystems can transform innovation into enterprises with global expansion capabilities. The critical question is no longer whether China is important to global innovation—that is already settled—but rather how global capital is reassessing China and which cities are most capable of translating this reassessment into tangible business creation, industrial capability upgrades, and genuine global influence. The answer lies within China's innovative landscape.

A new trend is emerging in the global venture capital market: while the scale of venture capital is recovering, capital is becoming more concentrated and selective. Global capital continues to flow, investors are still placing large bets, and new companies are constantly emerging. However, the market is no longer evenly distributed. Capital is increasingly funneling into a few core sectors, leading enterprises, and superior innovation ecosystems at an unprecedented rate.

Last year, global growth-stage venture financing rebounded significantly, reaching $425 billion. Half of this amount flowed to AI-related companies, with the United States alone attracting $274 billion, accounting for two-thirds of the global total. In the first quarter of this year, this concentration trend reached new heights, with record financing driven not by broad market recovery but by a few ultra-large-scale AI funding rounds.

Globally, the United States remains the largest market, with unparalleled depth in frontier artificial intelligence, software, and scalable capital. Europe's market is active but cautious, focusing more on later-stage investments. India continues to advance in digital and software sectors, with growing innovation momentum. China remains one of the world's most important and influential innovation markets.

With the proliferation and implementation of AI in Asia, driven by digital infrastructure and industrial technology, China's role in the global innovation landscape has become increasingly critical. The current trend is not a generalized return of capital to China but rather a reassessment of China through the lens of hard technology. In the previous cycle, the world viewed China largely through the grand narrative of consumer internet. Today, the lens is more focused and stringent. Investors are re-evaluating areas with undeniable capability barriers: AI models, robotics, advanced manufacturing, biotechnology, industrial software, and digital infrastructure. It is increasingly clear that in multiple strategically key areas, China's value extends beyond market size. More importantly, China is setting global benchmarks for cost-effectiveness and commercialization in certain leading fields, driven by strong competitiveness and astonishing development speed.

This does not mean China leads in all areas, that every sector receives equal attention, or that global capital is relaxing its scrutiny. On the contrary, today's investors are more discerning and focused than ever. Long-term capital holders are no longer pursuing broad macroeconomic factors but are increasingly concentrating on strategically significant sectors such as digital technology and foresight, advanced manufacturing and automation, and clean energy and green information. In other words, this is a selective reassessment centered on core strategic capabilities, not a simple return to past extensive investment models. This is why we must understand this transformation from the perspective of the innovation ecosystem.

A core viewpoint is that AI competition is first and foremost a competition for talent, followed by a competition of models. While it is often said that the AI race is about computing power, capital, and models, the foundation supporting all of this is always talent. Who is cultivating the top talent? Who is attracting this talent? Who is retaining them? And who is building the innovation ecosystem that allows them to truly start businesses and bring products to market? This is why talent data is so crucial. In the global AI talent landscape, a notable statistic is that 38% of top researchers at international premier AI conferences completed their undergraduate education in China. Additionally, many top AI talents educated in China are exerting influence globally. This reflects a highly balanced reality: China has established a world-class AI talent development system, and the mobility within the global innovation network allows these talents to create value on a broader stage.

Broadening the perspective reveals a more complete picture. China leads in AI paper publications, citations, and patent applications. In top-tier model R&D and private AI investment, the United States still maintains a lead, but the gap in frontier areas is narrowing rapidly. However, the competition is far from over. Understanding this race is not about simply determining who wins or loses; it is about two highly competitive innovation systems with distinct advantages pushing and advancing each other.

Global investors increasingly value not an isolated star city but a complete innovation system. Looking at China today, we see not just a single company, a wave of红利, or a fleeting slogan, but the rise of an entire innovation ecosystem. It is converging the most critical elements of this cycle: a continuous supply of talent, globally outstanding entrepreneurs with technical backgrounds, increasingly firm capital support and policy guarantees, and, most importantly, strong commercialization capabilities. The scale effect from cities to regions is also crucial. For example, in recent years, multiple domestic innovation hubs have continuously attracted a large number of young and high-end talents. The core digital economy industries contribute nearly 30% of GDP. These tangible figures turn the innovation ecosystem from a vision into an accessible reality.

We also clearly see that in frontier sectors such as large models, robotics, embodied intelligence, and spatial intelligence, a group of globally competitive enterprises is rapidly growing here. More importantly, the key conditions supporting these enterprises are maturing simultaneously. As global capital becomes increasingly selective, it is precisely such a complete ecosystem that will truly enter their field of vision.

The next focus of AI competition is not just model R&D but large-scale inference. Every AI call, every business process, and every device action consumes tokens. As models continue to optimize and inference costs rapidly decline, the economic value of AI is increasingly shifting toward large-scale deployment in real scenarios.

Therefore, the future of AI will depend less on whose demo is more impressive and more on who can make AI sufficiently cheap, reliable, and embeddable, truly integrating it into daily workflows, industrial systems, logistics networks, digital commerce, medical tools, and robotic devices. This is a completely different competition from merely training frontier models, and the underlying economic dynamics are changing at an astonishing speed. The cost per million tokens for a GPT-3.5 level system dropped from $20 at the end of 2022 to $0.07 by the end of 2024. Meanwhile, the domestic daily token call volume has exceeded 40 trillion.

In summary: Token costs are plummeting, while usage is exploding. In this competition, several key elements become crucial:普惠 computing power and compatibility, efficient and stable large-scale deployment capabilities, underlying power and infrastructure guarantees, and high-density, multi-scenario practical application落地. Application density is particularly critical. China's innovation hubs do not need to excel in every layer of the AI technology stack or meet all needs for everyone. However, in areas where China truly possesses advantages—AI applications, robotics, industrial intelligence, digital trade, and commercialization落地—it can build irreplaceable value. If the first phase of AI was about proving intelligence, the key to the next phase is embedding intelligence, truly integrating AI into the workflows of countless industries.

So, what specific opportunities will these trends translate into? We believe five major areas deserve attention:

First, AI application落地, where intelligence truly transforms into productivity. AI is reshaping enterprise processes, customer experiences, decision-making methods, and logistics systems.

Second, robotics and embodied intelligence. This is particularly important because AI is no longer confined to the digital world; it is increasingly entering the physical world,走进 factories, manufacturing, warehousing and logistics, healthcare, agricultural production, intelligent mobility, and service scenarios. This is no longer a niche sector.

Third, digital trade and cross-border infrastructure. This is one of the areas where China's digital trade ecosystem holds natural advantages. AI is reshaping every operational环节 of global commerce, from merchant tools and compliance systems to logistics coordination, multilingual translation, supply chain optimization, customer service, and trade empowerment.

Fourth, AI healthcare and biotechnology tools. The importance of this field lies in extending the innovation narrative from pure software to实体产业. AI is empowering scientific research productivity,辅助诊断, drug研发, medical system tools, and the transformation of scientific research成果.

Fifth, industrial technology and energy infrastructure. As AI is deployed on a large scale, efficiency itself will become a core strategy. The value of technology layers围绕 industrial and energy systems—such as software, sensing, optimization, battery management, control, detection, and predictive maintenance—will become increasingly prominent.

Finally, today's discussion can be summarized in three points:

1. Capital is becoming increasingly focused. Today's market rewards precise落地 rather than空泛 narratives, and it rewards innovation ecosystems that can demonstrate real capabilities. 2. Hard实力 is more important than ever. Talent reserves, industrial depth, and commercialization capabilities are becoming the core criteria for investors to judge where the next generation of industry-standard enterprises will emerge. 3. China's innovation ecosystem stands in an extremely advantageous position. Its value lies not in imitation but in constructing a unique combination: top-tier talent, an active innovative spirit, solid industrial support, deep platform基因, and强大的产业化落地 capabilities. Therefore, our opportunity lies not only in innovating here but also in starting from here to go global. We聚集人才, build products, and develop enterprises into companies with global influence. This is the core value of China's innovation ecosystem.

Thank you.