The core thesis is straightforward: a massive disconnect exists between China's contribution to the AI economy and the capital allocated to it. Goldman Sachs estimates that Chinese AI-related companies generate roughly 16% of global AI revenue, with an associated market capitalization of approximately USD 4 trillion, yet global mutual funds allocate only ~1.2% of their global technology portfolios to Chinese AI stocks .
Goldman argues that even a modest normalization of fund flows toward China's AI assets would drive significant relative outperformance . As one Goldman strategist put it, the thesis is not about any single AI application breakout but a "repricing opportunity driven by underweight positioning, policy investment, and hardware demand"
.
Goldman's report cites several concrete developments underpinning the bullish view:
1. Exploding chip exports. China's integrated circuit exports reached USD 43.3 billion in January–February 2026 alone, a 72.6% year-on-year surge . This reflects both domestic capacity expansion and AI-driven global demand. Earlier data showed China's IC exports hit a record $201.9 billion for all of 2025, up 26.8%
. The 111% surge figure sometimes cited likely refers to a broader period or a specific Goldman calculation; the best-verified SCMP and CNBC sources confirm a 73% increase in early 2026
.
2. Yangtze Memory Technologies' revenue explosion. The Chinese NAND flash maker saw Q1 2026 revenue soar ~445% year-on-year, capturing 13% of the global NAND market (tied for fourth place globally, up from 8% a year earlier) . The AI boom, robust domestic demand, and supply shortages fueled the surge
. Separately, DRAM maker CXMT posted 2025 revenue of approximately $8 billion, a 130% increase
.
3. Beijing's RMB 2 trillion data center plan. In June 2026, Bloomberg reported that China is preparing a five-year, RMB 2 trillion (~USD 295 billion) plan, led by the National Development and Reform Commission (NDRC), to build a nationwide network of data centers powered by at least 80% domestic chips . The following day, China's Ministry of Industry and Information Technology issued implementation guidelines for "AI + Information and Communication" development (2026–2028), signaling formal policy support
. Some analysts estimate the total investment could reach RMB 5 trillion (~$735 billion) when including power grid upgrades
.
The GSXACART launch sits within a broader rotation narrative. Goldman had held South Korean equities for over a year but now advises clients to shift capital from South Korean chip names toward the China AI basket . This follows a Goldman report from late June 2026 highlighting a divergence in the AI value chain: memory producers (including Samsung and SK Hynix) are extracting value from hyperscalers facing capex margin pressure
.
Separately, Goldman's China equity strategists have projected a 15% rise for the MSCI China Index in 2026, driven by earnings growth supported by AI, overseas expansion, and anti-involution policy . Goldman also noted that U.S. AI restrictions—the "Anthropic order"—have accelerated investor rotation into Chinese tech as differentiated exposure, with Chinese AI companies trading at significant discounts to U.S. counterparts
.
Goldman Sachs explicitly acknowledges several risks to this thesis:
The core figures—the 16%/1.2% allocation gap, GSXACART basket launch, RMB 2 trillion plan, YMTC 445% revenue growth, and 73% IC export surge—are consistently reported across multiple financial-news and brokerage platforms citing the Goldman report and official data . The precise "111% IC export surge" figure is not directly quoted in the highest-authority sources; the SCMP and CNBC sources show a 73% surge for early 2026, and the 111% number likely refers to a broader period or a specific Goldman calculation
. The 15% MSCI China Index target and South Korean chip rotation context come from earlier Goldman strategy notes rather than the July 9 thematic report itself
.