Chinese Hedge Funds Exit Nvidia — The AI Trade Shifts to Hardware Supply Chains

Perseverance Asset Management International slashed its position in Nvidia by 72% during the second quarter (Q2 2024), marking a decisive retreat by one of China's largest hedge funds (South China Morning Post Business).

Chinese Capital Pivots from Nvidia to Hardware Makers

The shift in Chinese hedge fund strategy represents a structural change in the AI investment playbook. Rather than betting on the dominant US-based GPU (Graphics Processing Unit, the specialized processor used for AI training and rendering) providers, these funds are aggressively migrating into the broader hardware supply chain (South China Morning Post Business). This move suggests that institutional investors in China are seeking exposure to the physical infrastructure of AI rather than the specific software-integrated silicon produced by US giants.

The scale of this reallocation is significant for global liquidity in the semiconductor sector. By reducing exposure to US hyperscalers (large-scale cloud service providers like Amazon or Microsoft) and Nvidia, Chinese capital is effectively re-routing the flow of liquidity toward manufacturers that support the entire AI ecosystem (South China Morning Post Business). This rotation could create a bifurcation in the semiconductor market, where US-centric chip stocks face selling pressure from overseas institutional outflows while hardware-centric manufacturers see increased inflows.

Nvidia vs. Hardware Supply Chain Makers

The divergence in fund activity highlights a growing gap between US-centric AI plays and the broader hardware ecosystem. While Nvidia has been the primary beneficiary of the AI gold rush, Chinese hedge funds are now prioritizing the components and manufacturing processes required to build these systems (South China Morning Post Business). This transition indicates a move from high-beta (a measure of a stock's volatility relative to the market) software-adjacent plays toward more diversified hardware infrastructure.

US-China Tech Tension Forces Localized AI Development

The necessity for localized AI models is driving a massive surge in custom chip development within the Chinese market. Apple has already begun training a custom large language model (LLM, a type of AI trained to understand and generate human-like text) specifically for the Chinese market with technical support from Alibaba Group (Reuters). This move is a direct response to the regulatory and geopolitical constraints that make standard US-based AI integrations difficult in the region.

This localization effort creates a massive new market for domestic semiconductor firms. As foreign companies like Apple seek to integrate local AI capabilities, they increase the demand for regional technical support and customized hardware solutions. This development provides a significant tailwind for Chinese-based chip designers who can navigate the complex local regulatory landscape (Reuters).

Supply Chain Constraints Threaten AI Scaling Speeds

The rapid expansion of AI capacity is facing a physical bottleneck: the availability of advanced chips. An investigation has highlighted an apparent discrepancy between the ambitious AI capacity goals announced by major tech companies and the actual number of advanced chips currently in operation (The Guardian Business). This shortage threatens to slow the deployment of large-scale AI models globally.

For investors, this scarcity creates a high-stakes environment for companies that control the chip manufacturing process. The inability to secure enough advanced silicon could lead to a slowdown in the projected growth of AI-driven services (The Guardian Business). If the supply of chips cannot meet the demand of the world's biggest technology companies, the entire AI investment thesis—which relies on the rapid scaling of compute power—could face its first major reality check.

IPO Activity Signals a New Wave of Hardware Funding

The capital requirements for scaling chip manufacturing are driving a new wave of public offerings in the region. Ingenic Semiconductor has launched a Hong Kong share offering to raise up to HK$3.22 billion (US$410.4 million), joining a growing trend of mainland Chinese chipmakers tapping capital markets to fund international expansion (South China Morning Post Business). This influx of capital is essential for firms looking to move beyond domestic markets and compete on a global stage.

This trend of chipmaker IPOs (Initial Public Offering, the process of offering shares of a private corporation to the public for the first time) suggests that the hardware side of the AI trade is entering a high-growth phase. As companies like Ingenic seek to fund their expansion, the semiconductor sector is seeing a renewed focus on manufacturing capacity rather than just design (South China Morning Post Business). This shift is critical for the long-term sustainability of the AI ecosystem, as the industry moves from the 'innovation' phase into the 'infrastructure' phase.

Key Developments to Watch

• NVDA (Ongoing) — monitoring the impact of Chinese institutional outflows on US-based semiconductor valuations
• Alibaba Group (by end of 2024) — progress on the localized LLM integration with Apple
• Ingenic Semiconductor (H2 2024) — the success of their HK$3.22 billion capital raise

As Chinese hedge funds pivot from Nvidia to the broader hardware supply chain, is the era of the 'ingle-stock AI winner' coming to an end in favor of a more fragmented, hardware-focused market?