Physical AI and Nvidia ETFs
Nvidia Leads the Era of Physical AI
From a Semiconductor Maker to an AI Platform Company
The highlight of CES 2025, the world’s largest consumer electronics and IT trade show, was Nvidia CEO Jensen Huang’s keynote address. The most closely watched segment was his roadmap outlining the evolution of artificial intelligence—from Generative AI, to Agent AI, and ultimately to Physical AI. Generative AI represents the current stage, epitomized by ChatGPT. Agent AI refers to service-oriented AI that functions as an assistant. Physical AI, the final stage, denotes humanoid robots equipped with AI “brains.” Nvidia’s message was clear: it is no longer merely a GPU supplier, but a provider of infrastructure across every stage of AI development.
The keynote opened with the announcement of Nvidia’s RTX 50 series GPUs. The mid-range RTX 5070 is expected to deliver performance comparable to the previous flagship RTX 4090 at a price of $549. Given that the RTX 4090 originally retailed at $1,599, the announcement immediately captured the attention of both the AI and gaming industries. Analysts expect the dramatic improvement in price-to-performance to drive explosive demand.
Market attention, however, quickly shifted to Cosmos, introduced as the infrastructure platform for Physical AI. Physical AI refers to artificial intelligence capable of learning and operating within the three-dimensional (3D) physical world, a critical requirement for autonomous driving and robotics. Together with Omniverse, Nvidia’s 3D simulation and training platform, Cosmos enables physical objects from the real world to be digitally replicated, virtual environments to be constructed, and countless scenarios to be tested and analyzed.
Just as Nvidia once used its CUDA (Compute Unified Device Architecture) software platform to lock developers into its GPU ecosystem, Cosmos is expected to play a similar role in autonomous driving and robotics—positioning Nvidia’s products at the center of industry advancement. In addition, the concept of “test-time scaling,” where AI dynamically adjusts computational resources based on the confidence of its responses, is also expected to further boost demand for Nvidia’s computing power.
Most ETFs that invest in the U.S. technology sector already include Nvidia as a core holding. Among them, three ETFs explicitly feature Nvidia in their product names: ACE Nvidia Value Chain Active (483320), ACE Nvidia Bond Balanced Bloomberg (448540), and TIGER Nvidia U.S. Treasury Covered Call Balance (Synthetic) (0000D0).
ACE Nvidia Value Chain Active is an equity ETF that invests in Nvidia and companies across the semiconductor value chain. ACE Nvidia Bond Balanced Bloomberg allocates 30% to Nvidia and 70% to domestic bonds, offering a more defensive structure. TIGER Nvidia U.S. Treasury Covered Call Balance combines 30% Nvidia exposure with 70% U.S. 30-year Treasury covered call strategies. In this structure, the covered call component dynamically adjusts option-selling exposure to target a monthly distribution of around 1%, blending participation in Nvidia’s price movements with income generation and stability.
Over the past year, Nvidia shares have risen 136.72%, while the five-year return stands at an extraordinary 2,159.79%. Debate over valuation peaks is inevitable. In particular, concerns have emerged as Nvidia’s gross margin has declined from a peak of 78.35% in April last year to 75.15% in July and 74.56% in October, raising questions about future volatility. Still, as highlighted in the CES keynote, the prevailing view is that Nvidia’s role will remain central throughout each phase of AI’s evolution.
For investors, the discussion ultimately comes down to how to gain exposure. With equity ETFs, bond balanced ETFs, and bond balanced covered call ETFs all now available, the choice depends on risk tolerance and investment horizon. Investors who expect long-term growth in the semiconductor value chain and can withstand short-term volatility may prefer equity-focused ETFs. Those seeking to retain Nvidia as a core long-term holding while partially cushioning volatility may find bond balanced or covered call ETFs to be more suitable investment tools.