Nvidia Order Slowdown Rumors Could Open the Door for AMD and Custom AI Chip Makers
Nvidia’s position at the center of the artificial intelligence hardware boom is facing fresh uncertainty as supply chain sources point to possible changes in the company’s AI chip ordering plans. The discussion comes amid growing attention on Nvidia’s China-related business, U.S. export controls, and reports that some China-focused H200 AI GPUs may be sitting unused in Hong Kong.
According to industry sources cited in Taiwanese media, Nvidia could slow certain AI chip orders as it reassesses the situation. If that happens, the impact may ripple across the semiconductor supply chain, particularly in areas where production capacity is already tight. The biggest potential winners could be AMD and companies developing custom AI chips, often known as ASICs.
Nvidia remains the dominant force in AI GPUs, with its accelerators powering many of the world’s largest data centers and AI training systems. However, its heavy exposure to China has become more complicated as U.S. regulators continue to tighten restrictions on advanced chip exports. Any increase in government scrutiny could force Nvidia to adjust its product strategy, shipment plans, or manufacturing commitments.
The controversy has also placed Nvidia CEO Jensen Huang under the spotlight. Reports have connected the situation to Nvidia’s relationship with China and to scrutiny involving a Singapore-based data center company. Nvidia has said it is conducting an internal investigation and has stated that no company secrets or national security concerns were compromised.
For the semiconductor industry, the bigger question is whether Nvidia’s situation could temporarily free up valuable manufacturing capacity. Advanced chip production is extremely limited, especially at leading foundries such as TSMC. Capacity for cutting-edge AI chips is planned years in advance, and customers typically make long-term commitments so foundries can justify expensive investments in new equipment.
If Nvidia reduces or delays some orders, even slightly, competitors may move quickly to secure that capacity. AMD is an obvious candidate. The company has been aggressively pushing its Instinct AI accelerators as an alternative to Nvidia’s data center GPUs. While Nvidia still leads in software ecosystem, market share, and performance perception, AMD has been gaining interest from cloud providers and enterprise customers looking for more supply options and lower costs.
Custom AI chip developers could also benefit. ASICs are designed for specific workloads rather than broad general-purpose computing. In artificial intelligence, this can make them more efficient and cost-effective for targeted tasks such as machine learning inference or large-scale model training. Google’s TPU is one of the best-known examples of this approach, and more major technology companies are exploring similar custom silicon strategies to reduce dependence on Nvidia’s expensive GPUs.
The potential opportunity is not limited to wafer production. Supply chain sources also point to components such as probe cards, which are used to test chips while they are still on silicon wafers. These tools are essential in semiconductor manufacturing, and supply can be limited. If Nvidia’s demand softens, even temporarily, ASIC developers and rival AI chip makers may be able to access resources that were previously difficult to obtain.
This matters because the AI hardware market is no longer driven only by peak performance. Cost, energy efficiency, availability, and supply chain flexibility are becoming just as important. Nvidia’s GPUs command premium prices because of their strong performance and mature software ecosystem, but many customers are looking for alternatives that can reduce costs at scale.
That shift could strengthen demand for ASICs. Custom AI chips may not replace Nvidia GPUs across the entire market, but they can work alongside them in data centers. For companies running enormous AI workloads, even small improvements in efficiency can translate into major savings. If export controls, supply constraints, or pricing concerns make Nvidia hardware harder to obtain, interest in ASICs and AMD accelerators could grow faster.
Still, a major displacement of Nvidia is unlikely in the near term. The company has built a powerful advantage through its hardware, networking products, software tools, and developer ecosystem. Data centers already optimized for Nvidia platforms may not switch quickly. However, any disruption in Nvidia’s order flow could give rivals a rare opening in a market where capacity is often locked up far in advance.
For now, the situation remains fluid. Nvidia’s AI chip business continues to benefit from massive global demand, but uncertainty around China sales and U.S. policy could influence how the company manages future orders. If supply chain reports prove accurate, AMD and custom AI chip designers may be positioned to capture newly available production capacity and strengthen their role in the rapidly expanding AI semiconductor market.
The AI chip race is entering a new phase. Nvidia still leads, but the combination of export restrictions, supply constraints, high GPU prices, and rising demand for customized AI hardware could accelerate diversification across the industry. For AMD, ASIC developers, and key semiconductor suppliers, any pause from Nvidia may become a valuable opportunity.
