After Jensen Huang Resigns, Taiwan Server Supply Chain Proactively Vets Customers

After NVIDIA co-founder Jensen Huang resigned from all positions, Taiwan server manufacturers are reportedly proactively strengthening customer and shipment-location checks amid concerns that the US may tighten export controls, with some even pausing orders from unfamiliar customers. Whether NVIDIA subsequently adjusts procurement, investment, and financing support will affect Taiwan’s chip manufacturing, packaging and testing, and server manufacturers.

According to reports, an executive at a Taiwan server company said he would personally travel to Southeast Asia to visit small customers and confirm whether the AI servers in their data centers are actually activated and operating normally. At his level, he did not need to conduct such checks in person in the past, but now he must be especially cautious to avoid violating the law.

An executive at another server manufacturer said the sales team has been required to suspend accepting orders from unfamiliar new customers and small customers, covering products from NVIDIA, AMD, and Intel. After all, supply is already tight, so the company is prioritizing large, trustworthy customers.

These actions reflect that even before the US announces new restrictions, manufacturers have tightened checks on their own. The impact may first affect order-taking and delivery procedures rather than factory capacity.

Hampden, senior vice president at Bain-owned supply chain advisory firm Proxima, believes that after Huang’s resignation, what truly affects supply is orders, long-term capacity commitments, sales channels, and regulatory changes. Feldman, a portfolio manager at ClearBridge Investments, also said NVIDIA’s product roadmap for the next few years is broadly clear; the issue lies in execution and scheduling.

Funding is another test. Huang supported AI labs and cloud providers’ expansion through equity investments, chip credits, and financing guarantees. If the new management becomes more conservative, customers’ arrangements for obtaining funding and expanding capacity may be affected, which would in turn affect demand for chips, advanced packaging, servers, and peripheral equipment. BofA analyst Arya believes that if NVIDIA becomes distracted or faces stricter scrutiny, AMD, Broadcom, and Marvell will benefit.

However, AI chip orders shifting to other design companies does not mean leaving the Asian supply chain. Hampden noted that Asia has deep foundations from advanced manufacturing and memory to packaging and testing, and can adjust with demand. Wafer manufacturing, advanced packaging, and high-bandwidth memory (HBM) capacity all have dedicated uses and cannot be switched to other products overnight. Global AI infrastructure demand remains strong, and a leadership change is unlikely to significantly slow data center expansion.

For Taiwan manufacturers, the key is to observe whether actual orders, delivery schedules, capacity commitments, and customer funding change, in order to judge whether the impact will remain at the level of market confidence or extend to revenue and investment.