Despite recent controversies over the profitability of AI (artificial intelligence), growth continues among U.S. semiconductor companies, including NVIDIA. This is because agent-based AI is on the rise and competition in ASICs (application-specific integrated circuits) is intensifying. In response, a new ETF (exchange-traded fund) investing in U.S. AI semiconductor stocks such as NVIDIA has been launched.
The '1Q NVIDIA Fixed US AI Semiconductor' ETF was listed on the 20th. This ETF focuses its investment on NVIDIA while also including companies that can benefit from the growth of the AI semiconductor industry.
What sets this ETF apart from other U.S. AI semiconductor ETFs is that it holds a small number of stocks—10—and fixes NVIDIA's weight at approximately 25%.
Regular rebalancing is conducted at the opening of the U.S. market on the first business day following the monthly Korean options expiration date, at which point NVIDIA's weight is adjusted to 25%. The maximum weight for any other stock in the portfolio is 20%. It includes TSMC, Broadcom, AMD, Micron, and others.
Ahn Sang-hyuk, Deputy General Manager of the 1Q ETF Solutions Division at Hana Asset Management, explained, "ETFs tracking the Philadelphia Semiconductor Index hold the top 30 semiconductor companies listed in the U.S." He added, "The 1Q NVIDIA Fixed US AI Semiconductor ETF holds 10 stocks to focus more on leading companies."
He continued, "NVIDIA's GPUs (graphics processing units) play a crucial role not only in AI training but also in inference." He stated that it is an ETF suitable for investors who want to clearly focus on NVIDIA and invest in major companies representing the AI semiconductor industry.
Although U.S. semiconductor stocks have been volatile due to recent controversies over AI profitability, Hana Asset Management still judges the growth potential of AI semiconductors to be high. This is because as AI model performance improves rapidly, demand is spreading from training to inference.
It is expected that Big Tech companies' demand for semiconductors will increase further as agentic AI—where systems understand goals and independently plan and execute multi-step tasks without human instruction—becomes fully realized.
Ahn explained, "As AI gradually enters the stages of practical application and commercialization, there is still room for additional growth in the U.S. AI semiconductor market." He noted that OpenAI's recently released 'GPT-6 Astara' and Meta's 'Muse' are examples illustrating this change.
He added, "Because actual market growth does not always correlate one-to-one with stock prices, there are times when prices fluctuate temporarily." He stated, "While volatility in semiconductor stock prices should be noted, the trend leading to training, inference, and agent-based AI has not been broken."
As the commercialization of AI gains momentum, market valuations are rising. According to Reuters, Anthropic, one of the representative AI companies, has a corporate valuation of around 2 trillion dollars (approximately 2,664 trillion won). This is double the valuation of approximately 1 trillion dollars (1,332 trillion won) as of the end of May last year.
The fact that Big Tech companies are developing ASICs is also acting positively for the U.S. semiconductor market. OpenAI has developed its own ASIC, 'Jalapeño,' in partnership with Broadcom. Anthropic is also forming an internal team to promote the development of semiconductors specialized for its AI models.
Ahn said, "Companies like Broadcom and Intel are facing new demand and collaboration opportunities that did not previously exist through growth in the AI semiconductor market and partnerships with Big Tech companies." He added, "The overall AI semiconductor market is expanding and gradually becoming more segmented from a GPU-centric market."
in STOCK
