President Trump Pushed for Nvidia China Deal — but It's AMD's Chips Beijing Wants

Nvidia won approval to sell H200 AI chips to 10 Chinese companies, including Alibaba, JD.com, ByteDance, and Lenovo, but Beijing then signaled tighter scrutiny of Nvidia GPUs, limiting the practical upside. The article argues AMD may be the cleaner China beneficiary after CEO Lisa Su reportedly met with China’s vice premier at his request and was invited to deepen cooperation. Nvidia remains dominant with roughly 80% to 90% AI GPU share and $215.9 billion in fiscal-year revenue, but China access remains politically fragile.

Analysis

The market is treating this as a binary win for the incumbent leader, but the second-order read is that Beijing is likely optimizing for leverage, not loyalty. If Chinese procurement starts favoring the vendor with less strategic visibility, the incremental prize may shift toward the “good enough” alternative rather than the market-share leader. That matters because in restricted markets, share often migrates to the supplier that creates the least policy friction, not the one with the best product.

For NVDA, the issue is not demand quality but demand durability. Any China-access headlines can lift near-term numbers, yet they also raise the probability of recurring approvals, scrubs, and administrative delays that make forecasting less reliable; that should compress the multiple if investors start discounting China as episodic rather than structural. The bigger hidden risk is channel inventory and qualification friction: if Chinese buyers hedge between domestic substitution and U.S. imports, NVDA could see lumpy orders without a corresponding improvement in long-term attach rates.

AMD’s opportunity is less about winning outright and more about becoming the politically tolerable second source. If that relationship deepens, the company can monetize China without carrying the same “must-contain” profile as the category leader, which is valuable in a market where regulators can shape purchase behavior. The catch is that this only matters if AMD can convert access into sustained design wins; otherwise, the rally becomes a policy premium, not an earnings premium.

Consensus is likely underestimating how much this dynamic benefits the rest of the AI supply chain. If China trims reliance on NVDA, the spillover can favor networking, memory, and foundry vendors that are less politically sensitive and more easily substituted across geographies. The trade is therefore not just long one name versus another; it is a relative bet on which companies can convert geopolitically fragmented demand into recurring revenue over the next 2-4 quarters.

Key Decisions for Investors

• Initiate a pair trade: long AMD / short NVDA on a 3-6 month horizon. Thesis is that AMD gets a policy scarcity premium while NVDA’s China optionality remains headline-driven; target 10-15% relative outperformance for AMD, with a stop if NVDA gets durable multi-quarter China order visibility.
• Buy AMD call spreads 3-6 months out rather than stock into weakness. The setup is asymmetric if Beijing continues signaling openness; use strikes that monetize a 15-20% upside move while capping premium bleed if the rhetoric reverses.
• Fade NVDA strength on China headlines by selling upside calls against existing longs or initiating covered calls for 1-2 quarter horizons. This expresses the view that China access is not yet a clean earnings bridge and the multiple already prices scarcity more than certainty.
• Add a basket long in non-NVDA AI infrastructure beneficiaries that can absorb redirected demand, especially names tied to memory, networking, and foundry exposure. The trade works if procurement in China becomes more fragmented, which would favor suppliers with lower geopolitical temperature.
• Set a catalyst watch for any formal China procurement guidance or export-control clarification over the next 30-90 days. If approvals become repeatable rather than one-off, rotate from policy trades into fundamental winners; if not, take profits quickly because the market will likely over-discount the transitory headline.