Nvidia and AMD face renewed China pressure as AI buyers plan to shift toward local chips

Nvidia and AMD are facing a fresh competitive threat in China as some AI buyers plan to allocate a bigger portion of their chip spend to local alternatives, according to a report cited by Yahoo Finance. The piece frames the shift as a near-term procurement change driven by China-based production plans, supply-chain preferences, and the realities of operating under evolving export and compliance constraints.

The article points to a survey outcome indicating that Chinese firms intend to reduce how much of their AI compute demand they satisfy with Nvidia and AMD chips. That direction matters for both companies because their data-center products, especially GPUs used for training and inference, have been central to how large-scale AI systems are built.

While GPUs remain a dominant approach for many AI workloads, the report’s premise is that buyers are increasingly willing to diversify hardware choices in China, including using domestically sourced accelerators. In general terms, AI hardware choices often differ not only by performance but by architecture. As industry overviews note, GPUs are general-purpose processors that can handle a wide range of training and inference tasks, while other accelerator types such as ASICs (application-specific integrated circuits) can be optimized for narrower AI functions, which can make them attractive when supply, integration, or cost priorities shift.

For Nvidia and AMD, the practical risk is that a procurement tilt away from their chips could translate into slower growth for the specific models and platforms they supply to Chinese customers. In addition to silicon itself, companies’ software ecosystems are typically part of the value proposition, but the Yahoo Finance account is focused primarily on buying plans and does not provide a detailed breakdown of which software changes, certification timelines, or performance thresholds buyers are using to justify switching.

Sector context adds to the sensitivity of the moment. AI chip competition has broadened beyond a single product category as governments and enterprises in key regions push for greater local manufacturing and resilience. That trend can compress the advantage of any one vendor if buyers determine they can meet their near-term needs with domestic solutions, even if those alternatives are not a perfect substitute everywhere.

It is also possible that the shift is uneven across industries and use cases. Some AI workloads are more tolerant of hardware variation than others, and buyers sometimes stage migrations, testing new chips in limited deployments before expanding. The report does not spell out whether the survey reflects immediate full-scale substitutions, phased pilots, or specific vertical categories such as consumer internet, cloud providers, or enterprise automation.

What remains unclear from the available reporting is how much spend is expected to move, over what time period, and which “local chips” the buyers are planning to use. The Yahoo Finance summary does not provide the survey’s sample size, the exact percentage-point allocation changes, or the identities of the local suppliers mentioned (if any), limiting how precisely investors and industry observers can translate the finding into revenue impact or share losses.

Going forward, the key developments to watch will be updates from China-focused buyers on procurement policies for AI accelerators, any public announcements from Chinese AI cloud and enterprise customers about partner hardware roadmaps, and indicates from GPU suppliers around China-related demand trends. Any quantified data, such as purchase-share targets, deployment scale, or confirmed model substitutions, would be particularly important for assessing whether this is a temporary adjustment or a longer-term structural change.