['New News'] 'Eating Meat and Drinking Soup': Huang Renxun's Capital Magic and the Unspoken Struggles of Taiwan's AI Supply Chain

Quanta's announcement of issuing 70 billion yuan in Global Depositary Shares (GDS) led to a stock price halt, revealing the harsh reality for Taiwan's AI server contract manufacturers: 'soaring revenue, declining gross profit, and strained finances.' However, if we zoom out to examine the entire global AI supply chain, this is not merely a financial issue specific to Quanta, but rather reflects an extremely asymmetric business reality across the global AI industry.

Huang Renxun is undoubtedly an exceptionally shrewd business genius. Leveraging his monopoly position in the high-end chip market, he has executed a sophisticated capital strategy: 'suppressing Taiwanese vendors' profit margins while extending payment terms.' This allows NVIDIA to capture the lion's share of value—the 'meat'—while leaving Taiwanese manufacturers with only the residual 'soup,' i.e., low-margin, high-volume manufacturing work.

In practice, despite Quanta's revenue increasing by 30% year-on-year in 2023, its operating margin dropped to just 4.2%. This indicates that even as the company secured significantly more AI server orders from NVIDIA, its profits failed to keep pace. More critically, NVIDIA imposes average payment terms of over 120 days on Taiwanese suppliers. As a result, manufacturers like Quanta deliver finished products but cannot receive payment for extended periods, severely deteriorating their working capital cycles.

This forces companies such as Quanta to rely on overseas fundraising instruments like GDS to supplement working capital for equipment investments and component procurement. This situation exposes the limitations of Taiwan's ODM model. While possessing world-class technical capabilities and mass production expertise, these firms lack brand power and ecosystem leadership, placing them at a constant disadvantage in pricing negotiations.

In high-growth markets like AI, first-mover advantage holders like NVIDIA maintain overwhelming dominance, often treating suppliers merely as 'factories.' Looking ahead, Taiwanese manufacturers may accelerate efforts to develop proprietary AI hardware architectures or offer integrated software-stack solutions to break free from structural dependency. However, such transitions require massive R&D investments and long-term strategic patience—challenges that remain formidable given many firms are currently focused on survival.