US Approves NVIDIA H200 Chip Sales to 10 Chinese Firms; Geopolitical Tailwind

The US government's approval of NVIDIA H200 chip exports to 10 Chinese companies marks a surprising policy shift from the Biden-era blanket semiconductor restrictions. Jensen Huang's appearance in Beijing alongside Trump signals tacit approval for selective technology transfers that maintain US competitive advantage while easing bilateral tensions. The H200, while a step forward for Chinese AI developers, is architecturally inferior to NVIDIA's H100 and does not represent a wholesale loosening of export controls on the most advanced silicon.

Market reaction was immediate and positive: NVDA rallied on the headline, extending its seven-day 20% run and pushing toward a $6 trillion market valuation. Investors interpreted the approval as a sign that the Trump administration may be willing to pursue selective engagement with China rather than total decoupling. This aligns with the broader Beijing summit narrative of managed trade relations. However, the practical impact on NVIDIA's financials is likely modest: the H200 is already being sold in North America and Europe, and the Chinese demand is not expected to materially move the needle on revenue given the chip's lower performance tier.

The geopolitical subtext is more important than the semiconductor economics. By allowing H200 exports, Trump signals that he views tech decoupling as a negotiating tool rather than a fixed policy. This could open the door to future semiconductor licensing agreements and potentially unlock billions in export revenue currently forfeited to policy restrictions. However, it also sets up a future flashpoint: if US-China relations deteriorate again, a rollback of these approvals could trigger equity market volatility and spark new tariff escalations.

Skeptics note that the H200 approval is being oversold by retail traders who conflate geopolitical relief with NVIDIA fundamental upside. The chip's technical limitations mean Chinese buyers will not be able to train cutting-edge large language models; instead, they will use H200s for inference and secondary workloads. If the narrative is that China gains meaningful AI parity via H200 access, that bullish case is overstated. Conversely, if tensions re-escalate and the US government reverses course on H200 approvals (as happened with other technologies), NVIDIA could face sudden supply cancellations and margin pressure.