Chinese Artificial Intelligence (AI) DeepSeek: ETF Winners & Losers

What is DeepSeek, and Why Should You Care?

For American AI companies, the emergence of DeepSeek is a wake-up call. While U.S. tech firms have poured billions into building state-of-the-art GPU infrastructure and massive data centers, the Chinese startup has managed to achieve world-class performance with far less capital.

The newly launched DeepSeek-R1 AI model has surpassed leading U.S. models like OpenAI's GPT-4 (internally referred to as "O1") and Meta's Llama across multiple benchmarks, showcasing better reasoning capabilities and efficiency.

The key innovation behind DeepSeek's breakthrough lies in its "distillation technique." Think of this as a process where a large, complex AI model teaches a smaller, leaner model how to think just as effectively.

By transferring advanced reasoning abilities from the bigger model to the smaller one, DeepSeek reduces the amount of computational power needed when users interact with it.

The result? A dramatic reduction in operating costs. For example, DeepSeek-V2, another version of the model, has an output cost of just $0.28 per million tokens compared to OpenAI O1's $60.00.

This cost difference is critical for applications like API access, where pricing for developers is significantly cheaper with DeepSeek.

Although DeepSeek still relies on tens of thousands of Nvidia H100 and H200 AI GPUs to train its models, China's AI engineers have developed key optimizations to work around U.S. export restrictions.

By finding ways to squeeze more performance out of limited hardware, DeepSeek has effectively minimized its dependency on the massive infrastructure investments that American companies rely on.

For investors, this raises concerns about how U.S. AI leaders like OpenAI, Meta, and others will compete if they face not only better technology but also cheaper operating models. It's no wonder that stocks tied to U.S. AI, particularly Nvidia, have taken a hit.

ETF losers

Unsurprisingly, thematic ETFs focused on AI were among the hardest hit. For instance, the Global X Robotics & Artificial Intelligence ETF (

), which has Nvidia as its top holding at 12.98% of its portfolio, dropped 3.91% on the day.

Semiconductor ETFs, particularly those heavily weighted toward Nvidia, also suffered. The VanEck Semiconductor ETF (

), which caps individual holdings at 20% and has Nvidia at that limit, fell 9.6%, reflecting the outsized influence of the company's slide.

Even broad market index ETFs were not spared. Due to the market-cap weighting methodology commonly used by these funds, which gives more influence to larger companies like Nvidia, the declines rippled through the broader market.

The SPDR S&P 500 ETF Trust (

) fell 1.78%, while the more tech-heavy and top-heavy Invesco QQQ ETF Trust (

QQQ

+1.91%

), which tracks the Nasdaq-100, dropped 3.13%.

This performance highlights one of the criticisms often raised by proponents of equal-weighted or fundamentally-weighted ETFs—namely, the concentration risk inherent in market-cap-weighted strategies.

The hardest-hit ETFs, however, were single-stock leveraged ETFs. For example, the GraniteShares 2x Long NVDA Daily ETF (

) magnified Nvidia's steep decline by two times, resulting in a staggering 31.23% loss on the day.

ETF winners

The sell-off wasn't universal across all sectors, which is why I remain personally optimistic that this isn't the end of the U.S. bull market—likely just a sector-specific correction.

Why? Defensive sectors showed resilience, with the Health Care Select Sector SPDR Fund (

) climbing 1.94% and the Consumer Staples Select Sector SPDR Fund (

XLP

-0.3%

) gaining 1.93%. These low-beta sectors often act as safe havens during times of uncertainty, providing stability amidst tech-driven volatility.

Remember Warren Buffett's famous quote: "When the tide goes out, you see who's been swimming naked." Well, Berkshire Hathaway (BRK.B), with its $325 billion cash pile, sat comfortably today, rising 1.5%.

While it's not an ETF, there is a newly launched ETF that tracks it—the Direxion Daily BRKB Bull 2X Shares (

)—which climbed 2.68% - not quite the 2x target though. Still, if there's any leveraged ETF I'd consider holding long term (though I generally don't recommend it), it would be this one.

This article is for informational purposes only and does not in any way constitute investment advice. The author may express their own opinions, which may not represent the opinions of ETF Central or its affiliated partners. It is essential that you seek advice from a registered financial professional prior to making any investment decisions.