Nvidia’s Physical AI Shift Sends Asian Costs to 90% and Amazon Eyes $50B Chip Sales
Nvidia’s expansion into physical AI has raised Asian suppliers’ share of production costs to 90% from 65%, driving double-digit gains at LG Electronics and regional partners. Amazon is weighing external Graviton and Trainium chip sales that could generate $50 billion and threaten Nvidia’s data center GPUs.
1. Nvidia’s Physical AI Expansion
Nvidia has broadened its AI focus beyond semiconductors into robotics, autonomous systems and AI-enabled manufacturing. This strategic push has increased Asian suppliers’ share of production costs from 65% last year to 90%, underscoring the region’s critical role in the company’s evolving ecosystem.
2. Double-Digit Gains for Asian Partners
Following Nvidia’s physical AI announcement, LG Electronics shares jumped 15% on plans to integrate home robots with Nvidia’s platform, while Nanya Technology climbed 10%. Huizhou Desay SV Automotive and Pateo Connect also saw significant stock uplifts after unveiling intelligent driving collaborations with the chip designer.
3. Hyperscaler Spending Drives Growth
U.S. hyperscalers Amazon, Microsoft and Alphabet are each committing nearly $200 billion to AI capital expenditure this year. Nvidia accounts for almost half of Microsoft’s capex and a quarter of Amazon’s, creating substantial downstream demand for hardware suppliers like Hon Hai and SK Hynix.
4. Amazon’s $50B External Chip Ambition
Amazon is planning to offer its custom Graviton and Trainium chips to third parties, projecting external sales could reach $50 billion annually if production capacity exceeds AWS requirements. This move would position Amazon as a direct competitor to Nvidia in the data center GPU market.