Large-Cap Growth Outlook: Guarded Optimism but Volatility Is Likely

Key Takeaways

Our outlook for large-cap growth companies over the coming months is cautiously optimistic. We believe significant innovation and broader access to artificial intelligence (AI) could spur earnings growth and stock prices in 2025. However, high valuations and the market’s narrowness have left many large companies susceptible to volatility amid geopolitical turbulence and domestic economic and policy uncertainty.

We think high-quality businesses with experienced, agile management teams will be best positioned to navigate this environment. Since not all companies will fare well, we believe this could be an opportunity for active stock selection to shine.

Market Concentration: How We Reached This Point

It’s important to highlight how narrow the market’s performance has become. In 2024, the 20 largest stocks in the Russell 1000^® Growth Index accounted for 85% of the index’s return, according to FactSet. In contrast, the remaining companies contributed only 15% of the index’s return.

In addition, after two years of gains, the market has become quite expensive.

Simply put, the markets have started 2025 on a high, narrow perch.

We believe this analogy is helpful because it highlights the market’s potential downside risk. Recent bouts of volatility reinforce the point. They include the December 2024 stock sell-off and the U.S. tech stock sell-off in January after Chinese AI developer DeepSeek released its R-1 open-source large language model.

Exploring Innovations Beyond AI in 2025

While the media and markets have been fixated on AI, we see tremendous innovation and opportunity throughout the U.S. economy. For example, we expect corporations to continue investing in business continuity and supply chain redundancy initiatives, helping fuel a stronger and more resilient economy.

We also expect continued investment in enterprise digital transformation, which should bring greater data-driven insights, better decision-making, higher agility and more productivity. We also think cybersecurity will be an area of innovation and focus in 2025 and beyond.

The Trump administration appears to support greater investment in technology, infrastructure spending and onshore manufacturing. Finally, a healthier financial sector and the potential for a sustained period of lower interest rates would increase the availability of capital for growth investments.

Beyond Mega-Cap Tech: Market Performance Outlook

The mega-cap stock rally began with the public launch of ChatGPT in November 2022. Since then, the focus has been on AI infrastructure, including chip makers, data center operators, cloud providers, and energy companies that support creating, training and maintaining large, generative AI (GenAI) models.

Two years later, some of this infrastructure is in place, opening the door for AI's efficiencies to spread throughout the economy. Even before the DeepSeek announcement, we saw new GenAI capabilities and applications that promise increased productivity and potential breakthroughs across many industries.

We now expect even more companies to start seeing the productivity benefits of this technology. We anticipate a faster shift from developing AI models to using them to generate real-world insights. For example, Nvidia reported in August that more than 40% of its data center revenue now comes from using AI for practical applications, indicating that this transition is already underway.¹

DeepSeek’s Impact on GenAI Accessibility

The DeepSeek model's unveiling suggests that smaller, less costly models can be developed quickly, reducing the need for extensive training. Additionally, it claims to have been created using older Nvidia chips. If true, this lowers the barriers for companies to adopt generative AI and raises questions about the necessity of the newest, most innovative chips.

We consider this a major source of uncertainty to monitor. In January, the Trump Administration announced a $500 billion public/private data center project, and Microsoft and Meta Platforms have set hefty AI spending targets.

Here’s the worry: If effective, low-cost open-source models can be developed, do we need massive data centers filled with the latest chips?

The U.S. approach to GenAI stands on three legs — model size, computing power and data access. DeepSeek threatens to knock out the first two legs of the stool, making it a significant risk to the AI trade.

Investors and corporate leaders need to determine the right amount of AI infrastructure. If a small start-up like DeepSeek can achieve cutting-edge results at a fraction of the cost, being a model vendor may not offer a competitive advantage. It is hard to justify massive AI investments when others can easily replicate the latest models.

Because DeepSeek is an open-source model, its innovations will spread to other model vendors, reducing costs and making these tools more accessible. As a result, one of DeepSeek’s biggest impacts might be on the revenue potential of companies developing and monetizing closed, proprietary AI models.

Another big DeepSeek contribution could be that it shows the viability of smaller models, enabling powerful GenAI tools to run locally on consumer devices.

Meeting the Rising Demand for AI Infrastructure

Perhaps the most intriguing issue with DeepSeek is its impact on AI infrastructure spending. We believe smaller, cheaper and more efficient AI models will boost adoption and usage, leading to a greater need for infrastructure. This is known as the “Jevons Paradox,” where more efficient use of resources increases overall resource consumption. This makes sense because high computing and training costs are barriers to AI entry, and DeepSeek effectively removes these barriers.

As AI adoption and applications continue to grow, further investment in physical and tech infrastructure should be considered. This means we see ongoing opportunities not only in semiconductors, data centers and power but also in digital security and domain-specific expertise to maximize AI’s impact. In 2025, the focus is likely to shift more toward the application layer of AI technology.

But to be clear, AI advancements won’t benefit all companies equally. This isn’t a corporate arms race to infinity and beyond. Eventually, AI investments must translate into revenue and profit growth. Some companies will do this better than others. As a result, we believe investing in AI requires a careful, company-by-company approach.

Looking back with 25 years of hindsight at the dot-com experience, we can see that the internet wasn’t a universal salve for all corporate wounds. Instead, some companies seized the opportunity and monetized the technology radically better than others. We believe we’re in a similar situation with AI today. It’s still a new technology, and we’re waiting to see if it can drive productivity and profits beyond advertising spending.

Evaluating High-Quality Companies Amid Economic Uncertainty

We conclude that markets will likely be volatile and prone to short, sharp moves throughout 2025. We expect market movements to be uneven but generally trend upward. This environment helps support the case for active stock selection.

Regardless of economic and market conditions, owning great businesses over time remains our North Star. We believe investing in well-managed, competitively advantaged companies that can sustain growth and profitability may offer the best way to navigate this uncertain environment.