The AI Investment Shift: Beyond Chips, Towards Software

 The Data Layer as the Next Growth Engine:

Companies positioned between storage and applications—specifically those with consumption-based models like Snowflake, MongoDB, Datadog, and Elastic—are primed to capture the massive surge in AI data processing. Models are useless without clean, maneuverable data.

The End of "SaaS-mageddon":

The broad-stroke selloff in software is transitioning back into a stock-picker's market. High valuations (e.g., Snowflake's 112x forward P/E or Datadog's 98x) are being justified by strong annual revenue growth (33% and 25%, respectively) driven by enterprise consumption demand.

The Rise of Orchestration and Routing:

As AI enters an "agentic" phase, efficiency and cost-control are paramount. Private companies like Not Diamond, Martian, and OpenRouter are creating value at the application level by dynamically routing workloads to the most cost-effective models, a software parallel to the complex coordination happening at the chip design level.

To conclude, the thesis presented here offers a logical framework for optimizing the Cloud Software segment of a technology-heavy portfolio. While the semiconductor super-cycle and the dominance of Hyperscalers laid the foundation of the AI boom, the data layer represents the necessary infrastructure for sustained, day-to-day enterprise adoption. Balancing allocations in Chips and Hyperscalers with high-growth, consumption-driven software companies provides targeted exposure to this exact operational phase of the AI rollout.

To discuss these implication further, you can reach me, Sean Wagman, at 416-733-5660.