AI、 The Future of Finance and Capital Markets | Key Points of Gary Gensler's Speech Productivity and Growth What truly drives GDP is not simple automation, but transformation and enhancement. Automation itself may not necessarily drive growth, the key is how market concentration, employment, wages, inflation, and financial stability evolve. At some point in the future (perhaps in the 2030s), AI will trigger a financial crisis. Not necessarily as severe as in 2008, but the risks are real. The Difference between Optimism and Prudence Top economists have divergent opinions: some believe that there will be no significant productivity increase in AI within the next decade, while others are extremely optimistic. Reality may be somewhere in between. AI is essentially a nonlinear, non deterministic, and hyper dimensional mathematical system, with explanatory difficulties that are difficult to solve in the short term. Concentration and Geopolitics The high concentration of chip manufacturing, rare earth refining, TSMC and other links has brought about a game between countries. Manufacturers rely on "moats" to lock in users and strengthen dependence. Data leakage and price opacity are also hidden dangers, especially in the financial industry. The competition between China and the United States is the main axis, and how Europe, Latin America, Asia, and Africa can maintain resilience in "computing power and data dependence" is a key challenge. Reflection of the Capital Market In 1964, the top four companies by market value were IBM, GM, AT&T, and ExxonMobil. Today, the top 10 companies account for 40% of the market value of the US stock market, with several individual companies exceeding trillions of dollars. The total market value of the US stock market is equivalent to 225% of GDP, accounting for about half of the global stock market. Only four giants (Amazon, Google, Microsoft, Meta) invest $300 billion annually in AI CapEx, almost equivalent to 1% of GDP. Half of the US GDP growth in the first half of the year came from spending on chips and data centers. Financial Applications and Risks Successful implementation: customer service, claims documentation, sentiment analysis, small-scale quantitative transaction feature extraction. • Limitations: High inference latency and error rate, not yet suitable for high-frequency trading. The financial industry relies on outsourcing and service providers, and small and medium-sized financial institutions need AI service providers, which is a great opportunity for startups. However, it is necessary to pay attention to risks such as fraud, manipulation, and deepfakes. Regulation needs to add "barriers" to AI to prevent financial fraud and market manipulation. Foam and Difference • The AI investment boom is different from the railway foam in the 19th century and the real estate foam in 2008: most of the funds come from giants with abundant cash flow, rather than debt. • So even if there is a foam, its nature is more controllable. Advice for Entrepreneurs 1. Service positioning: Provide external services such as data cleaning, model orchestration, RAG, etc. for financial institutions. 2. Compliance framework: Long term development must be within the scope of legal policies and avoid pure speculation. 3. Defend customers and data: Banks have huge profit margins, and the entry point for fintech is to find data and customer insights. 4. Internal Disruption: Allow "entrepreneurs" within the team to drive AI applications, otherwise they are prone to external disruption. Comparison of Cryptocurrencies The value proposition of AI is clear, while Crypto mostly remains focused on emotions and speculation. Stablecoins are variants of the US dollar, and the long-term value of other tokens is questionable. The capital market will ultimately allow fundamentals to catch up with emotions. Conclusion AI will continue to be a transformative force, but accompanied by concentration, financial crises, geopolitical and regulatory challenges. For enterprises and investors, opportunities lie in deep application, vertical services, and compliance innovation. For a country, how to establish resilience in data, computing power, and regulation will determine its future on the global AI map.