The universe, ranked

Sorted by AI Leverage Score — a transparent blend of exposure purity, input scarcity, conviction tier and time-to-revenue. Click any row for the full company view. Every column header sorts.

The stack, from the AI dollar outward

Barclays’ 19 subcategories are not a flat list — they are a causal chain. A dollar of hyperscaler capex enters at Layer 1 and works its way down to the molecule. The further down you go, the further from the AI narrative you sit, and the closer you get to the physical constraint that actually gates the buildout. Click any subcategory to filter the universe.

The asymmetry worth noticing: the layers with the most companies are not the layers with the tightest constraints. Attention pools at the top of the stack, where the story is cleanest; the binding shortages — turbine slots, large power transformers, and licensed electrical crews — sit two and three layers down, where the supplier count is thinnest and pricing power is most durable.

Purity against scarcity

The two questions that decide whether a name on this list is an AI investment or an AI anecdote. Horizontal: how much of the company is actually levered to the buildout. Vertical: how scarce what it sells is, relative to demand. Bubble area is market capitalisation; colour is stack layer. The filters above apply here too.

Method & sources

What came from Barclays, what is our own judgement, and where the seams are. Read this before you act on anything above.