Hong Kong's AI Ambition: A Hub in Search of a Conscience

The numbers are staggering, and they demand our attention. Over the past six months, AI-related initial public offerings in Hong Kong have raised nearly HK$100 billion, accounting for 55% of all new listings on the exchange. Let that sink in for a moment. More than half of the capital flowing through one of the world's premier financial arteries is now tagged with the artificial intelligence label. As someone who has spent years watching capital markets dance with technological narratives, I can tell you this: that kind of concentration is not a signal of health. It is a signal of fever. And in a fever, we must be the ones who keep our heads clear while others are seeing visions. This is not a story about whether AI is real. It is real. The technology is transforming industries, and Hong Kong's Financial Secretary, Paul Chan, has made it clear that the city intends to ride this wave with a comprehensive government push. The AI Efficiency Group has already launched 30 projects across 13 departments. The export sector is seeing double-digit growth, driven by global demand for AI hardware. The narrative is one of momentum, opportunity, and a city positioning itself as the world's application hub for artificial intelligence. But beneath this glossy surface, there are structural questions that no policy speech will answer. And as a community that has weathered the ICO mania of 2017 and the DeFi summer of 2020, we have a responsibility to look deeper. Let me give you some context from my own experience. In 2017, I was the lead community liaison for MakerDAO's early development team here in Cape Town. I watched as 500 speculative tokens flooded the market, each one promising the moon while delivering nothing but empty smart contracts. I organized town halls, vetted submissions, and tried to protect non-technical investors from catastrophic losses. What I learned in those trenches is that the gap between technological promise and practical reality is where both fortunes and tragedies are made. Hong Kong's current AI push has that same energy, that same intoxicating blend of policy support and capital enthusiasm. And it carries the same risks. The core of Hong Kong's strategy is what I would call "application-led, efficiency-first." The government is not trying to build foundational models. It is not competing with Beijing, Shenzhen, or Hangzhou in the race to create the next GPT. Instead, it is taking mature technologies and adapting them to government workflows, financial services, and public administration. This is a rational choice, given the city's resource constraints. Hong Kong lacks the deep research institutions and massive compute clusters that fuel foundational model development. Its path to value creation lies in scenario adaptation and systems integration. But here is the uncomfortable truth that the policy documents do not mention: this strategy makes Hong Kong a permanent follower in the AI technology stack. It will be dependent on external model suppliers, whether that is Alibaba's Qwen, DeepSeek, or the American giants like OpenAI and Anthropic. And in this dependency, there is a quiet vulnerability that no amount of IPO volume can mask. I have audited enough projects to know that when a government or corporation claims to be "AI-powered," the reality is often far more mundane. The 30 efficiency projects across 13 departments sound impressive, but they likely involve document processing, data analysis, and public service chatbots. These are valuable use cases, to be sure. But they are engineering-level innovations, not architectural breakthroughs. They are the application layer, not the foundation. And while the application layer can create significant economic value, it does not confer strategic independence. The question that keeps me up at night is not whether these projects will succeed. It is whether the people implementing them understand the difference between using AI and owning the means of AI production. Now, let me address the elephant in the room: the 55% IPO concentration. I have seen this movie before. In 2000, it was internet companies. In 2017, it was ICOs. In 2021, it was NFTs. The pattern is always the same. Capital floods into a narrative, valuations detach from fundamentals, and then the reckoning comes. The Hang Seng Index has started including AI companies, which will attract passive fund flows and further inflate valuations. But here is my concern: how many of these "AI companies" are genuinely building core technology, and how many are simply traditional businesses with an AI label slapped on their pitch deck? The definition of "AI-related" in these statistics is dangerously broad. It includes fintech companies using machine learning for credit scoring, logistics firms optimizing routes with algorithms, and a host of others where AI is a feature, not the core product. This is not necessarily a problem, but it is a risk. When the market corrects, and it always corrects, the distinction between real AI value and narrative-driven speculation will become brutally clear. The 650 billion Hong Kong dollar opportunity is another figure that deserves scrutiny. The report suggests that if small and medium enterprises could match large enterprises in AI adoption by 2035, the economic benefit would be substantial. This is a compelling vision, but it is a potential value, not a guaranteed outcome. The gap between SME and large enterprise AI adoption exists for structural reasons: cost, talent, infrastructure, and awareness. Closing that gap requires more than policy statements. It requires subsidies, training programs, and a fundamental shift in how small businesses perceive technology. I have seen this challenge firsthand in my work with women in emerging markets through the SoulBound cooperative. We onboarded 1,500 users in 2020, and I can tell you that the barriers to adoption are not technical. They are educational, cultural, and financial. The same will be true for Hong Kong's SMEs. Let me also raise a point that the official narrative conveniently ignores: the compute infrastructure gap. Hong Kong has not announced any major AI computing center plans. It has not addressed the physical constraints of land scarcity, high energy costs, and a climate that is hostile to data center operations. The city will likely rely on cloud services from Alibaba, Tencent, or AWS, or it will tap into the computing resources of the Greater Bay Area. This creates a dependency that has both supply chain and data sovereignty implications. When government AI applications involve sensitive citizen data, the question of where that data is processed and stored becomes not just a technical issue but a political one. The absence of a sovereign compute strategy is a strategic blind spot that could undermine the entire application-led approach. Now, let me offer a contrarian perspective. Perhaps Hong Kong's "application-first" strategy is not a weakness but a strength. In a world where foundational model development is increasingly concentrated in a few global players, the real value creation may shift to those who can effectively deploy and adapt these models to specific contexts. Hong Kong's unique position as a bridge between mainland China and the global market, its common law system, its international professional services ecosystem, and its free flow of information give it a distinctive advantage. The city can be the testing ground where AI applications are validated for the Asian market, the regulatory sandbox where cross-border data solutions are developed, and the financial hub where AI companies access capital. This is the "super-connector" role, and it has genuine value. The question is whether this role is sustainable in the face of competition from Singapore, which is investing heavily in AI research, talent, and infrastructure. I have been through enough market cycles to know that the current enthusiasm will eventually meet reality. The question is not whether there will be a correction, but how severe it will be and who will be hurt. The 55% concentration of AI-related IPOs is a warning sign. It suggests that capital is chasing a narrative rather than fundamentals. When the correction comes, the "fake AI" companies will be exposed, and the entire sector's credibility will suffer. This is why I believe the government must establish clearer standards for what qualifies as an AI company, requiring disclosure of core technology capabilities and the percentage of revenue derived from AI products. This is not about being anti-AI. It is about protecting the ecosystem from the damage that comes from inflated expectations and subsequent disillusionment. There is also the human dimension that the policy documents overlook. The government's push for AI efficiency will inevitably change the employment landscape. Administrative positions in the public sector will face automation pressure. The financial services industry, which accounts for a significant portion of Hong Kong's GDP, will see job transformation. The article does not mention any comprehensive retraining programs or social safety nets for workers displaced by AI adoption. This is a critical omission. In my 2022 bear market series, "Stoicism in the Bear Market," I emphasized the importance of emotional resilience and community support during times of crisis. The same principle applies here. Technological transitions are not just technical challenges; they are human ones. A society that embraces AI without addressing the human cost is building on sand. Let me also raise the ethical dimension, which is entirely absent from the official narrative. Government AI applications involve citizen data, and the question of algorithmic transparency is paramount. Do citizens have the right to know when AI is used in decisions that affect them? Is there a mechanism for independent audit of government AI systems? What about algorithmic bias? If a government AI system has biases, it could create systematic unfairness for specific groups. Hong Kong has not yet enacted comprehensive AI regulation, and the "one country, two systems" framework creates a complex compliance environment. The city must navigate between mainland China's AI regulations and international standards like the EU AI Act. This is not an easy balance, but it is a necessary one. Code is law, but ethics is conscience. And without a clear ethical framework, the application of AI in government could erode public trust in ways that are difficult to repair. I want to be clear about what I am not saying. I am not saying that Hong Kong should abandon its AI ambitions. The strategy of leveraging mature technologies for economic benefit is sound. The focus on application and integration is appropriate for the city's resources and strengths. What I am saying is that the current approach has blind spots that need to be addressed. The compute infrastructure gap needs a plan. The talent shortage needs a comprehensive strategy that goes beyond rhetoric. The SME adoption challenge needs concrete programs, not just aspirational targets. And the ethical framework needs to be developed in parallel with the technology deployment, not after the fact. In my work with the Ethereum Foundation on the Human-Centric AI whitepaper, I learned that the most successful technological deployments are those that keep human values at the center. We drafted guidelines to ensure that AI-driven DAOs remain accountable to human oversight. The same principle applies to government AI applications. Technology must serve human dignity, not the other way around. This is not a naive idealism; it is a practical necessity. Systems that do not earn public trust will fail, regardless of their technical sophistication. As I look at Hong Kong's AI push, I see both promise and peril. The promise is real: a city leveraging its unique position to become a global AI application hub, creating economic value and improving public services. The peril is equally real: a bubble in AI-related valuations, a talent shortage that constrains growth, a compute infrastructure gap that creates dependency, and an ethical framework that lags behind deployment. The path forward requires honesty about these challenges and a commitment to addressing them with the same energy that is being applied to the technology itself. Solidarity over speculation. This is the principle that has guided my work through bull markets and bear markets, through technological hype and disillusionment. It is the principle that I believe Hong Kong needs to embrace as it navigates this AI transformation. The city's success will not be measured by the amount of capital raised or the number of AI projects launched. It will be measured by whether the benefits of AI are distributed broadly across society, whether the human costs are addressed with compassion, and whether the ethical foundations are as solid as the technological ones. Culture on-chain, heart on-screen. The technology is a tool, but the values are the foundation. And in the end, it is the values that will determine whether this ambitious push becomes a sustainable transformation or just another chapter in the history of technological bubbles. The next 18 months will be telling. Will we see concrete results from the 30 government projects? Will there be a comprehensive talent strategy? Will there be an announcement about compute infrastructure? Will the SME adoption gap begin to close? These are the signals I will be watching. And I would encourage everyone in this ecosystem to watch them as well, not with the feverish excitement of speculation, but with the clear-eyed focus of those who understand that building something lasting requires more than just riding a wave. It requires building the infrastructure, the human capacity, and the ethical framework to sustain it. Hong Kong has the opportunity to be a model for how a city can embrace AI while maintaining its values. The question is whether it will seize that opportunity or squander it in the pursuit of short-term gains. The answer, as always, lies in the choices we make today.