Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Nvidia Holds 90 Percent of the GPU Market as a Japanese Samsung Distributor Picks Rebellions' NPU

Nvidia is reported to hold about 90 percent of the GPU market, a dominance that remains hard to challenge directly. Yet a Japanese distributor of Samsung products has chosen the rival Rebellions NPU, suggesting alternatives are beginning to win real deployment decisions in AI acceleration.

Published
报道:英伟达占GPU市场约90%,日本三星分销商转向Rebellions NPU
Image source: nvidia.com

A report from Tech Times offers two snapshots of the AI chip market. On one side, Nvidia is said to hold around 90 percent of the GPU market. On the other, at the level of actual procurement, a Japanese distributor of Samsung products has chosen a rival NPU from Rebellions.

The 90 percent figure reflects ecosystem inertia as much as silicon. GPUs come bundled with software stacks, framework support, developer habits and machine supply chains, which makes displacement slow even when a competitor's hardware looks attractive.

The distributor's choice points to a different dynamic. When buyers start seriously evaluating non-Nvidia options, competition stops being purely about benchmark performance and becomes a mix of supply, cost and local support.

NPUs are typically designed for inference workloads, where efficiency per watt can matter more than raw throughput, which is why they keep coming up in edge and on-premises deployments.

For a challenger like Rebellions, winning a place in a distributor's catalogue can matter more than a launch-day spec sheet, because distribution is where designs turn into deployed systems.

What to watch is whether the arrangement grows into larger deployments, whether Nvidia can defend its roughly 90 percent share, and whether more regional distributors begin treating NPUs as a standard alternative to GPUs.

Why it matters

A single distribution decision will not dent Nvidia's roughly 90 percent share, but it shows NPUs moving from fallback to viable option in inference procurement. If more such choices follow, the AI chip contest shifts from peak performance toward supply and local ecosystems.

NvidiaRebellionsGPUNPU
Back to realtime news

Nearby Updates

All