Wall Street heavyweight Goldman Sachs has issued a major research report urging global institutional investors to systematically go long on China’s artificial intelligence value chain. The report, which departs from conventional rating language to explicitly recommend a “Long China AI Value Chain” trade, highlights a severe structural undervaluation of the Chinese AI sector relative to its actual industrial output and global revenue contribution.
According to the analysis, detailed by BigGo Finance, the core rationale for this strategic pivot rests on a compelling “triple mismatch.” As of early 2026, China’s AI sector possessed a total market capitalization of approximately $4 trillion, representing only 10 percent of the global AI market cap. However, the sector contributes 16 percent of global AI-related revenue. Meanwhile, global mutual funds and large asset managers allocate a mere 1.2 percent of their portfolios to Chinese equities.
Goldman estimates that even a modest convergence of global capital allocation toward the revenue share would unleash incremental capital inflows exceeding $100 billion. The firm explicitly states that there is no AI bubble in China; rather, the potential economic benefits of the sector are 50 to 100 percent above current market pricing. This assessment stands in stark contrast to other global markets, where AI valuations have frequently outpaced fundamental earnings delivery.
The Divergence Between Hard Tech and Soft Tech
Within the Chinese equity landscape, the Goldman report identifies a historic divergence in performance between A-share “hard tech” companies and H-share internet platforms. The STAR50 index, which represents onshore AI hardware and infrastructure firms, has outperformed the offshore, internet-heavy Hang Seng Tech Index (HSTECH) by a staggering 68 percentage points year to date.
This divergence reflects the differing positions these companies occupy within the AI value chain. Hard tech firms, including semiconductor manufacturers, equipment suppliers, and infrastructure builders, are direct beneficiaries of the massive capital expenditure required to build AI capabilities. In contrast, soft tech and internet platforms are currently bearing the brunt of these costs.
The internet sector has accumulated over 180 billion yuan (approximately $26.6 billion) in subsidy losses since the second quarter of 2025, as companies aggressively discount their AI services to capture market share.
“We reiterate our broad tactical preference for A-shares over H-shares, and hard tech over soft tech,” noted Kinger Lau, Goldman Sachs Asia portfolio strategist. However, Lau added that after significant valuation compression, select H-share internet names have entered an observation zone, potentially pivoting toward a profit-recovery trade if earnings rebound in the coming quarters. This recovery will depend heavily on whether these platforms can successfully monetize new AI opportunities, such as AI shopping agents and enterprise cloud services.
Global Capital Rotation and Hardware Explosions
The Goldman report arrives amid a broader realignment of global AI capital. In the first half of 2026, significant investment was concentrated in South Korean and Taiwanese markets. However, concerns over valuation froth and the sustainability of AI capex have triggered a rapid rebalancing. On July 8, South Korea’s KOSPI index plunged approximately 5.4 percent, officially entering a technical bear market, as foreign investors withdrew a net total of more than $100 billion year to date.
As capital rotates out of crowded trades, it is increasingly flowing into China’s core AI sectors. This influx is supported by undeniable industrial momentum, particularly in the semiconductor space. Despite intense US export controls, China has developed a comprehensive, full-value-chain competitive advantage spanning equipment, materials, wafer fabrication, and AI chips.
The explosive growth of domestic memory leaders exemplifies this trend. Chinese DRAM manufacturer CXMT has successfully passed the Shanghai Stock Exchange’s IPO review and plans to list on the STAR Market, aiming to raise billions in what could be the largest A-share IPO in years. The company’s 2026 revenue is expected to surpass $50 billion, securing its position as the fourth-largest DRAM manufacturer globally. Similarly, YMTC reported a 445 percent year-on-year surge in revenue, expanding its global NAND flash market share from 8 percent to 13 percent.
A 1.2 Trillion Yuan Industry
The scale of China’s AI ambitions is further contextualized by data released ahead of the 2026 World Artificial Intelligence Conference (WAIC) in Shanghai. According to official previews, China’s core AI industry surpassed 1.2 trillion yuan ($165.78 billion) in scale by the end of 2025, encompassing more than 6,200 enterprises.
This massive industrial base is driving rapid advancements across the technology spectrum. The ASPI Critical Technology Tracker recently noted that China now leads in 57 of 64 critical technologies, underscoring the depth of the country’s research and development capabilities. The rapid maturation of China’s embodied AI sector and the deployment of advanced foundational models demonstrate that this industrial scale is translating into tangible technological leadership.
For global investors, the Goldman Sachs report serves as a stark reminder of the disconnect between geopolitical narratives and market realities. While international headlines often focus on the constraints imposed by US sanctions, the underlying data reveal a Chinese AI ecosystem that is not only surviving but rapidly expanding its global market share. As the industry continues to mature, the trillion-dollar revaluation opportunity identified by Goldman Sachs may prove increasingly difficult for global capital to ignore.