Datt, who focuses on ASX small cap investing, says the AI-driven data centre buildout has created a multi-year construction cycle that is reshaping capital allocation across the local market — but that the attention concentrated on data centre operators and real estate investment trusts is obscuring a more compelling opportunity one level down the value chain.

"The companies attracting the most attention, the data centre owners and operators, are not necessarily where margin concentrates," Datt said.

The ASX is not the US semiconductor trade

Datt draws a sharp distinction between how global investors are playing AI infrastructure and what is actually available on the ASX.

In the United States, semiconductor and memory companies have captured the largest share of AI-driven investment returns. The ASX offers no direct equivalent to those names, but it does provide exposure to the physical layer that underpins data centre construction and fit-out — and Datt argues that distinction matters enormously for how local investors should position themselves.

"Global AI adoption has driven sustained demand for data processing capacity that is reshaping capital allocation patterns across the ASX," he said. "In the United States, semiconductor and memory companies have been the primary beneficiaries of this shift, but the ASX does not offer direct equivalents. What it does offer is exposure to the physical infrastructure layer that supports data centre construction and fit-out."

"Australian investors accessing this theme through ASX-listed names are not buying the same trade as US semiconductor investors, and understanding that distinction clearly is the starting point for sound positioning in Australian small cap equities," Datt said.

Three tiers, three different propositions

To help investors navigate the theme, Datt outlines three distinct categories of ASX exposure to the data centre buildout, each with a different risk and return profile.

"Investors have to understand exactly where they're investing in the value chain. The first category is holding the data centre itself, digital real estate investment trusts. The second is construction, the direct contractors and electrical specialists. The third is equipment supply, the distributors servicing these centres," he said.

Data centre REITs and asset owners, the first category, operate on yield-driven return logic with long development cycles and meaningful sensitivity to interest rate movements. Datt places them in a structurally different bucket from the construction and supply tier — and it is the latter two categories where his focus sits.

He applies a framework familiar to resources investors to explain the logic.

"The picks-and-shovels framework is well understood in resources investing, and the same logic applies to the ASX data centre theme," Datt said. "When a sector experiences a surge in capital deployment, the businesses supplying the inputs often capture more consistent margin than those building or holding the end asset, because their competitive position rests on specialist capability rather than balance sheet scale."

Pricing power, he argues, is a key feature of the contractor and distributor tier. "Data centre construction requires specialist electrical contractors, fit-out expertise, and technology distribution networks that are not easily commoditised, and the constrained supply of qualified operators at this layer creates pricing power that feeds directly into earnings," he said.

Small cap derating creates entry points

Datt also points to current market conditions as a reason why small cap exposure to this theme may be particularly timely. He notes that index-level stability on the ASX is masking significant stress in smaller names, with many small and mid-cap stocks having derated materially through May and June 2026 — including businesses with sound underlying fundamentals.

"The current ASX environment masks significant divergence beneath index-level stability, with concentration in the top 20 stocks creating an appearance of calm that does not reflect conditions across the broader market," he said.

"Many small and mid-cap names have derated materially through May and June 2026, including businesses with sound fundamentals and no deterioration in their underlying earnings trajectory, as investor risk appetite has narrowed toward liquidity and certainty at the large cap end."

For Datt, that kind of broad-based selling is precisely when small cap investing becomes most productive. "This divergence is the condition that makes ASX small cap investing most productive, because sector-wide selling that compresses valuations indiscriminately creates entry points in businesses whose competitive position remains intact," he said.

He describes the data centre adjacent names as sitting squarely within this dynamic, with the thematic well recognised at the index level but the specific opportunity at the small cap tier requiring more granular, category-level analysis to identify.

Category selection over sector selection

Datt's broader point is that for investors seeking to build exposure to AI infrastructure through the ASX, choosing the right part of the value chain matters more than simply selecting the right sector.

"The construction and equipment supply tier offers a more defensible margin profile, lower capital intensity, and a more direct link between activity levels and earnings than the asset ownership layer, making it a more disciplined vehicle for capturing thematic tailwinds," he said.