- A core-satellite portfolio splits investments into stable core holdings and higher-risk satellite picks.
- The core is usually 60% of the portfolio, with satellites at 40%.
- It blends passive index investing with active opportunity bets.
The deal to sell Nvidia's AI chips to China was supposed to be one of the big stories of 2026. It's turning into a slow-motion stall.
Trump reversed years of policy last December, greenlighting Nvidia to sell its H200 chip — a powerful AI accelerator — to Chinese customers. The market cheered. Alibaba, ByteDance, and Tencent were already lining up to buy. Nvidia had forecasted potential sales worth tens of billions.
Then things got complicated.
The Financial Times reported that Nvidia has quietly stopped manufacturing H200 chips earmarked for China, betting that ongoing regulatory friction will keep the market effectively closed. And this week, Bloomberg reported that the Trump administration is weighing a new restriction: capping H200 exports at 75,000 chips per Chinese customer. AMD's comparable chip would count toward the same limit.
Seventy-five thousand chips sounds like a lot. It isn't — not for the scale of what China's tech giants were hoping to build. The South China Morning Post reported that companies like Alibaba and ByteDance privately told Nvidia they wanted far more than that. The 75,000 limit is less than half of what either firm requested.
Total exports to China could theoretically reach one million units — but only across all Chinese customers combined, and only if the cap doesn't change again.
Meanwhile, Nvidia CFO Colette Kress warned investors that Chinese competitors "are making progress and have the potential to disrupt the structure of the global AI industry over the long term." Huawei's AI chips, once considered a distant second, are getting closer to competitive.
China was 13% of Nvidia's revenue in fiscal 2025 — and that was with restrictions already in place. A full reopening could have added billions more. Now Nvidia isn't even building inventory for it.
The good news: Nvidia doesn't need China to grow. Revenue surged 73% last quarter to $68 billion. Demand from U.S. and European data centers remains massive. The bad news: every month the China window stays closed is another month Chinese AI companies spend building alternatives.
Nvidia called China an upside — not a necessity. At this rate, it might stay that way.
