今天寫的：Where is the "Scarcity" in the Age of AI?

Where is the Scarcity in the Age of AI? ??
When analyzing deep, long-term tech trends, Baillie Gifford is an indispensable reference. Known for their 10-year research horizons, they look beyond short-term market noise to uncover non-consensus correctness and non-linear growth.

In their latest global research agenda, a core pillar stands out: The Lens of Scarcity.

Every technological revolution shifts value by turning historical scarcity into abundance, while creating brand-new bottlenecks. Those who control the bottlenecks become the next corporate titans.

Here is how Baillie Giffords Global Alpha team breaks down the shifting landscape of AI scarcity:

1?? The Rise of Physical Scarcity ??

While software becomes abundant, the physical infrastructure to run AI faces acute scarcity. US hyperscalers are projected to spend nearly $700B in capex in 2026 alone.

Storage (HBM): High Bandwidth Memory is in critically short supply. Buyers are locking in 3-to-5-year contracts to secure inventory. Agentic AI will only accelerate this memory demand.

Processing: Beyond GPUs, the rise of Edge AI (running models locally on phones, cars, and robots) will spark massive demand for mass-market CPUs, ASICs, and power semiconductors.

The Energy Bottleneck: Compute demands are testing global grids. By 2030, global data center power demand is projected to more than doublematching Japans total electricity generation in 2024. Companies modernizing power grids hold incredible defensive value.

2?? The End of Scarcity in Software? ?

Conventional wisdom warns that enterprise software face major disintermediation risk. If an AI agent can build, maintain, and optimize code via simple plugins, traditional software moats risk eroding. Past disruptors risk becoming the disrupted.

3?? Where Will AI Moats Endure???

According to Baillie Giffords AI Defensibility Matrix, the safest growth opportunities lie in high scarcity and low disintermediation risk:

Trust Accuracy: Verified data and security will command a premium (e.g., Mastercard, Adyen).
Physical Advantages: Companies seamlessly blending digital AI with physical assets, logistics, or factory automation (e.g., Samsara, Keyence).

Adaptability: Platforms pivoting to become gatekeepers of online content, ensuring creators are compensated by web-scraping models (e.g., Cloudflare).

As long-term investors, navigating AI requires cognitive flexibility. The physical layers of the value chain show immense resilience, while software layers will see a wider dispersion of winners and losers.