China has banned Nvidia’s RTX 5090D V2 gaming chip during high-level diplomatic talks between President Trump and Chinese leadership, the Financial Times reported. The chip, originally designed for the Chinese gaming market but widely used by AI developers, was quietly added to a customs list of prohibited goods while Nvidia CEO Jensen Huang was in the country for the summit. The timing raises tensions in an already strained tech relationship, as Huang had just expressed optimism about future market access.
The RTX 5090D V2 was Nvidia’s second attempt to comply with US export controls for the Chinese market, reducing memory from 32GB to 24GB to limit AI capabilities. Despite these cuts, reports emerged that AI companies were modifying the cards to double memory to 48GB, likely drawing Beijing’s scrutiny. China’s ban effectively kills the chip’s primary market, with retailers like JD.com already removing special “AI GPU” sections that had listed the product.
In parallel, Alibaba announced its latest AI chip, the Zhenwu M890, at its Cloud Summit. The chip is three times faster than its predecessor and optimized for complex AI agent software. Alibaba also unveiled two future chips and a 128-accelerator server system, highlighting Beijing’s push to build domestic alternatives to Nvidia. T-Head, Alibaba’s chip unit, has shipped over 560,000 Zhenwu chips to 400+ customers across car manufacturing and finance.
Nvidia faces additional headwinds. Gamers criticized the company’s new AI-driven graphics system for over‑processing visuals, while the chip ban adds uncertainty ahead of Nvidia’s fiscal Q2 earnings, where analysts expect $1.77 EPS on $78.97 billion revenue. Meanwhile, the US has eased some AI chip exports to China, allowing orders of up to 75,000 H200 units for select firms — a contrast to China’s own blocking of the gaming-focused RTX 5090D V2.
