These are the key 2026 investing themes to watch

By Fox Business Clips

Share:

Key Concepts

  • Data Centers: Expanding infrastructure crucial for AI and computing power.
  • Space Economy: Growth driven by defense budgets, broadband, and emerging data center applications in space.
  • Compute Dynamics: Shift towards competitive chip market with focus on TPUs, UWS chips, and lightweight configurations.
  • Cost Per Token: A key metric for AI affordability and adoption, currently high but expected to decline.
  • Networking & Optics: Critical components for connecting GPUs and enabling data transfer, especially in space-based data centers.
  • AI Agents: Limited adoption currently due to high cost per token.
  • Bitcoin Mining & Data Center Infrastructure: Transition of companies like Applied Digital towards building custom data center infrastructure.

Investing Themes for 2026 & Beyond: A Deep Dive

Ivana Delefska, Founder of Speer Investment, outlined her firm’s investing themes for the coming year, emphasizing the evolving landscape of technology and its investment opportunities. The discussion, led by Charles, centered around data centers, space, compute dynamics, and the cost of AI.

Data Centers & Space: A Synergistic Opportunity

Delefska asserts that both data centers and space are “a reality,” despite appearing futuristic. She anticipates companies preparing for launches in 2027, projecting space to become a “sizable market.” A key bottleneck to realizing this potential is connectivity, specifically networking. Significant investment is being directed towards understanding how GPU’s (Graphics Processing Units) connect, both terrestrially and in the context of space-based data centers. This necessitates advancements in optics and networking components.

The upside for the space economy is threefold: increasing defense budgets allocating significant funds to space, a growing broadband market (SpaceX exceeding $20 billion, with another $10 billion+ for the rest of the value chain), and the emerging opportunity of data centers in space. Delefska believes the latter will provide the “real upside opportunity,” building upon a solid base case provided by the first two factors.

Compute Dynamics & the Race for Efficiency

The compute space is predicted to become increasingly competitive in 2026 and 2027. The emergence of TPU’s (Tensor Processing Units) next year, alongside the UWS chip, signals a trend towards “lightweight and different chip configurations” for data centers. Weight is identified as a critical factor, driving innovation in chip design.

Delefska differentiated between NVIDIA’s infrastructure for general-purpose computing (the Kuta system) and the underlying hardware (TPUs, CPUs, GPUs), characterizing the Kuta system as more of a “software layer.”

The Cost Per Token: A Barrier to AI Adoption

A central theme is the cost per token, defined as the expense of utilizing Artificial Intelligence. Currently, this cost remains “very high,” hindering the widespread adoption of applications like AI agents. Companies are unable to offer these services for free due to these high costs.

The question of whether demand drives price down or vice versa was raised. Delefska argues that companies must invest in “higher-end chips,” and a delay in NVIDIA’s deployment of the H100 Ultra has contributed to the sustained high cost per token. She anticipates “significant declines” in cost with the deployment of the H100 Ultra.

Stock Picks & Investment Strategies

Delefska highlighted three specific stocks:

  • Rocket Lab: Favored due to the aforementioned drivers in the space economy. The stock has performed well over the past three years, benefiting from defense spending and broadband expansion.
  • A-Lab: A long-term holding, despite its volatility. A-Lab is well-positioned in the networking and connectivity space, providing unique components with high market share. The volatility stems from speculation regarding its inclusion in specific chip designs, presenting a buying opportunity.
  • Applied Digital: Initially a “juggernaut” in 2020, the stock is regaining momentum. The company is transitioning from Bitcoin mining to building data center infrastructure. Unlike Neo Cloud (which builds and sells as a service), Applied Digital secures contracts to build infrastructure specifically for customers. Delefska believes this model is superior and could potentially lead to a conversion to a REIT (Real Estate Investment Trust), resulting in a higher valuation multiple.

Logical Connections & Supporting Evidence

The discussion demonstrates a clear connection between advancements in computing power, the need for efficient networking, and the expansion of data center infrastructure – both on Earth and in space. The high cost per token is presented as a limiting factor for AI adoption, directly linked to the availability of advanced chips like those from NVIDIA. The stock picks are presented as direct plays on these trends, with specific details provided to support the investment thesis. Delefska’s arguments are supported by market data (SpaceX exceeding $20 billion in revenue) and observations about industry dynamics (the competitive landscape of chip manufacturing).

Synthesis & Main Takeaways

The core takeaway is that the convergence of AI, space exploration, and advanced computing presents significant investment opportunities. While challenges remain – particularly regarding the cost of AI and the complexities of space-based infrastructure – the long-term outlook is positive. Investors should focus on companies positioned to benefit from these trends, particularly those involved in networking, chip manufacturing, and the development of specialized data center solutions. The transition of companies like Applied Digital from Bitcoin mining to data center infrastructure highlights the evolving nature of the technology landscape and the potential for disruptive innovation.

Chat with this Video

AI-Powered

Load the transcript when you're ready to chat so the initial page stays lighter.

Ready to summarize another video?

Summarize YouTube Video