Initial take on AI selloff and DeepSeek

What happened 

Several U.S. technology firms that have been at the forefront of the artificial intelligence (AI) revolution and areas leveraged to AI, such as data centers and utilities, are under pressure today. DeepSeek, a Chinese AI startup, recently introduced its AI model that shows competitive performance with some of the dominant U.S. tech models at a fraction of the cost with what the company has described as less-advanced chips.

This development has created uncertainty about whether the massive capital expenditures by U.S. companies is justified and whether competition will erode their robust profit margins and earnings. Moreover, stocks that were at the forefront of producing high-performance semiconductors are being hit the hardest.

Our Take

From a tech perspective, it’s pulled forward some of the discussion around sustainability of spending, returns on invested capital, and how much end use demand exists.

Our take is AI, with its productivity benefits, is in a very nascent stage. The amount of data that will be needed to support its growth is still tremendous and we have only scratched the surface. This will continue to require cutting edge technology.

However, not everything needs the same computing power and there is a benefit for more efficient and less expensive models. That is, cheaper models and chips will allow greater usage across industries and the economy. 

Already, while some of the big tech companies are selling off today, we are seeing rotation into software stocks. They potentially benefit from cheaper access to AI tools, making it more economical to incorporate into use cases.

Indeed, as Satya Nadella, Microsoft’s CEO, commented: “Jevons paradox strikes again! As AI gets more efficient and accessible, we will see its use skyrocket, turning it into a commodity we just can't get enough of.” Indeed, as a corollary, the creation of the smartphone exploded the use of the internet.

Longer term, there has not been a point in history where computing power and demand have decreased. There is an open question on what this news means for profit margins of specific companies, such as the Magnificent 7.

Can DeepSeek derail AI-related CapEx spending?

It’s doubtful that a Chinese startup will dominate globally given the current geopolitical tensions between China and the U.S.

Chinese smartphones, for example, are generally comparable and sometimes superior in terms of technology, battery life, and computing capability, yet they remain underutilized by Western consumers. Similarly, Chinese electric vehicles (EVs) surpass their Western counterparts in design, mileage, and cost of ownership but are still largely unavailable in the U.S. and other Western countries, where efforts to limit their presence persist. DeepSeek might face a similar fate.

Importantly, at the database level, backend applications depend heavily on support from Oracle and Microsoft, as these companies dominate global database structures. They are unlikely to embrace open-source applications, as they rarely support such developments.

Corporate users of AI primarily rely on widely adopted applications like Microsoft 365 and Google Apps. Beyond this space, the use of alternative applications will likely remain limited.

Risks of disruption

If DeepSeek’s claims are true—that it offers a robust AI tool utilizing low-level Nvidia chips, open-source code, and can be deployed at a fraction of the cost—then it could inspire hundreds of copycats, especially in the U.S. In such a scenario, large-scale capital expenditure projects will need to prove their commercial profitability. This potential disruption is the risk the market started pricing in this week.

Positioning & bottom line

A key theme in our outlook has been to stick with the primary uptrend but expect disruptions along the way. The disruption caused by DeepSeek was from a less obvious corner of the market. Indeed, with so much focus on Washington, this news is a reminder that other factors tend to matter more in totality for markets.

We maintain a positive long-term view of the U.S. market. There are open questions regarding pricing power for many of the megacap tech stocks and at what level they will spend. Given hefty gains for many of the AI-related stocks over recent years, the repair process from this selloff will take time. Based on what is known currently, we expect AI demand to remain firm and we remain positive on tech longer-term.

We are mostly seeing money rotating within the market as opposed to large outflows today. Although AI related stocks are down sharply, the S&P 500 Equal-Weight Index, a proxy for the average stock, is down less than 0.5%, and four sectors are trading higher. If there is a silver lining, the rotation underway could lead to somewhat more of a balanced near-term market.