Big Tech Accumulates Heavy AI Infrastructure Assets

The Lowy Institute reports that the ongoing AI buildout is shifting leading technology firms away from an asset-light model toward heavy physical infrastructure, citing investments in data centres, energy access, semiconductor supply chains and physical compute capacity. The article states that capex is expected to reach $2 trillion soon, more than double year-ago levels, and that fixed assets already account for about one half of Big Tech. The piece frames AI compute as a quasi-utility input tied to industrial policy, energy security and geopolitical competition.

Companies making comparable transitions historically confront multi-year capital cycles and new operational failure modes. These patterns typically raise the importance of energy procurement, thermal management, hardware lifecycle planning, and site-level redundancy for teams that previously focused on software-only risks.

Industry observers increasingly treat large-scale compute as strategic infrastructure rather than a transient R&D expense. This elevates questions around cross-border supply chains for semiconductors, long-term power contracts, and the macroeconomic implications of concentrated capital spending in a handful of firms.

Track disclosed capex and fixed-asset growth in company filings, announcements about long-term energy deals or new data centre regions, and policy moves that link critical infrastructure to national security. For practitioners, monitor operational metrics that reflect hardware reliability and energy intensity rather than purely software deployment KPIs.