The 32nd Annual Meeting of The Chinese Finance Association (TCFA) was held from November 7 to 8 under the theme "Investing in the AI Era." Sina Finance participated in the annual meeting as a media partner.
On November 8, during the "AI and the Global Economy II" roundtable discussion held at Fordham University in New York, Marc Giannoni, Chief U.S. Economist at Barclays; Mark Patrick, Managing Director and Head of Macro, Country Risk and Credit Risk at TIAA; and Shen Jianguang, Vice President and Chief Economist of JD.com Group, jointly explored AI investment, productivity, inflation, and U.S.-China competition. The roundtable was moderated by Henry Mo, Chief Economist and Head of Market Strategy at Corebridge Financial.
Regarding the recent repricing of AI-related assets, Giannoni argued that a distinction should be drawn between financial market valuation adjustments and the real investment cycle. AI infrastructure construction will not stop in the short term, but even if investment continues to grow, its contribution to economic growth may decline as the pace of investment growth slows. By his estimate, AI construction will directly contribute about 0.6 to 0.7 percentage points to U.S. GDP growth in 2026. A moderate cooling would help make the pace of construction more sustainable.
Turning to productivity, Giannoni stressed that building up the capital stock and improving the efficiency of capital use are two separate stages. At present, overall U.S. data has not yet fully shown that AI has delivered broad-based total factor productivity gains by transforming corporate processes and creating more output with the same inputs. Recent productivity growth has also been driven by increased capital input and higher labor utilization intensity. He is optimistic about AI's future role but believes macro evidence is still needed.
Shen Jianguang, by contrast, argued that AI's boost to the economy can already be observed from a broader perspective. Citing the more dynamic performance of China and the United States compared with Europe and Japan, he emphasized the support that AI investment, technological development, and wealth effects provide to growth and consumption. Giannoni responded that he does not deny these growth contributions, but that investment and wealth effects should still be distinguished from improvements in overall production efficiency.
From China's perspective, Shen Jianguang said that data as a production factor, the digital economy layout, and power infrastructure provide support for AI development. He believes the United States has advantages in frontier models and chips, while China has favorable conditions in application costs and power supply. However, rapid supply-side development does not mean demand-side problems will resolve themselves naturally. He stressed that China still needs to expand consumption and service demand, shift more resources from physical infrastructure investment toward human capital and livelihood-related areas, and implement the transition from investment to consumption.
Patrick's core concern is the possible time gap between investment expansion, productivity realization, and cash returns. He said that technology ultimately transforming the economy does not mean the financing process will not experience asset declines or financial shocks. Taking data center investment as an example, he values reliable leases, secured power supply, and controllable construction risks, and is wary of deteriorating deal quality after large inflows of capital. He pointed out that if actual returns fail to keep pace with investment, a sharp asset price adjustment could hit the economy, but this is a risk scenario that needs to be assessed in advance, not his baseline forecast. On the other hand, even if AI brings significant productivity growth, if the gains are overly concentrated among a few capital owners, it could weaken social support for technological development.
On inflation and monetary policy, Giannoni said AI may expand supply capacity over the long term, but currently strong investment and consumption demand alongside relatively insufficient savings will put upward pressure on the equilibrium interest rate and may keep short-term inflation pressure in place. Patrick added that amid higher uncertainty, investors should not rely too heavily on a single baseline scenario.
As for whether China and the United States are heading toward different AI ecosystems, Patrick said the two sides' supply chains are highly intertwined, making it difficult to reduce dependence; national security, political constraints, and key links such as chips could all change corporate investment and financing conditions. Shen Jianguang emphasized that beyond competition, there is still a need for cooperation, and many areas are difficult to fully separate.
In his summary, Henry Mo noted that it is necessary to separately discuss whether AI is technologically transformative, when it will manifest as productivity and economic growth, and whether current investment can generate sufficient financial returns. "These three outcomes do not necessarily arrive at the same time."
Sina Statement: All meeting transcripts are compiled from on-site stenographic records and have not been reviewed by the speakers. Sina publishes this article for the purpose of conveying more information and does not mean it agrees with or confirms the views and descriptions expressed.
Comments