Can DeepSeek's AI Disrupt Markets and Create Opportunities?

What’s the story?

Stocks tied to the artificial intelligence (AI) theme experienced increased volatility starting on Monday, 27^th January, after investors feared disruptions from Chinese AI startup DeekSeek. The open-source startup offering a similar chatbot app to OpenAI’s chatgpt is sending shockwaves as it claims its AI model is much more cost-effective, outperforming US competitors. Since then, Chinese tech giant Alibaba has also launched its own new AI model claiming it can beat DeepSeek.

In other words, the global race to produce a better and better AI model is heating up and productions costs are becoming a focal point, forcing the industry's biggest players to re-evaluate their strategies.

Why it matters and what does it mean for investors?

The first conclusions are that low-cost models should be positive for AI adoption. In essence, it echoes Jevon’s paradox, where technological progress through DeepSeek’s cost-effective models could boost overall AI demand, thereby increasing the need for AI equipment and services. For that reason, we remain overall positive on AI stocks, but there are nuances to which stocks should benefit most from this technological advancement.

AI models are set to continue evolving, becoming more powerful and moving AI towards mass adoption, in our view. Technological progress is a strong positive.

Figure 1. AI models are growing 10x per year, or 100x every two years.

Source: Google, BNP Paribas Exane

Revisiting the AI ecosystem

The move towards more powerful large language models (LLMs) requires innovation across computing, networking, and system memory to strengthen power. It also places more emphasis on networking infrastructure to support the increasingly complex network of compute nodes.

The expanding AI supply chain, particularly in networking, presents significant growth opportunities, in our view. In that sense our specialists have identified companies that could be the key winners of the AI ecosystem.

Figure 2: Key Enablers of the AI Semis and IT Hardware Component Supply Chain

Source: BNP Paribas Exane

Does it change forecasts of CAPEX for Hyperscalers?

The reduced costs for LLMs raise questions about whether hyperscalers (companies offering massive cloud services) will need to maintain their high level of spending on AI infrastructure. The next few weeks may provide hyperscalers the opportunity to reevaluate, but in our view, most of the spending should remain, as building AI infrastructure is crucial to maintaining their business moats and competitive advantage. Additionally, faster AI adoption could also boost demand for AI infrastructure. Therefore, it may be a bit early to determine the exact impact of DeepSeek on hyperscalers CAPEX.

What are the implications for electricity demand?

The trends associated with the need for more data processing and storage should continue, in our view. Moreover, the accelerating democratization of AI could lead to broader demand for data center capacity in the US and abroad. Finally, the trend towards electrification extends far beyond the expected additional demand from data centers, particularly in industrial decarbonization, but also in transportation (increasing EV demand) and potentially in heating (heat pumps).

Investing takeaway

Investors should stay diversified and use market volatility as an opportunity to invest in quality companies within the growing AI ecosystem. The advancements in AI, like DeepSeek's cost-effective models, offer both opportunities and risks. Diversification helps manage these risks while capturing growth potential.

Leveraging expert views and structured strategies can optimize risk-reward outcomes. Staying informed about key AI enablers and industry trends will help investors capitalize on the evolving AI market effectively.

This article is brought to you by the Advisory Solutions Team.