Macquarie Asset Management
The latest research and outlooks in the ProAllocate library
- Thematics
From prompts to power: Generative AI and implications for data centre demand
Macquarie Asset Management models generative AI's electricity use from actual usage patterns rather than chip shipments, projecting a ninefold rise to 857 TWh by 2040 requiring up to $5.6 trillion in infrastructure capital. It argues inference will dominate demand, making power delivery timing and location the critical investment question.
- Private Markets
Core infrastructure: The case for a long-term strategic allocation
Macquarie Asset Management argues core infrastructure merits a permanent portfolio allocation, citing a 9.9% annualised return since 2008 with bond-like volatility, inflation resilience and a 4.0% yield exceeding equities and bonds.
- Private Markets
Powered land is becoming the industry’s scarcest asset
Macquarie Asset Management's James Kemp argues that value creation in Asia-Pacific data centres has shifted upstream to securing powered land rather than operating buildings. AI training demand is driving gigawatt-scale campuses requiring new capital structures, power sourcing strategies and investor capabilities.
- Private Markets
Infrastructure secondaries: the specialist advantage in a complex market
Infrastructure secondaries volumes have grown 28% annually since 2016 to roughly $20bn in 2025, with pricing far less volatile than buyout or real estate secondaries. Limited dry powder relative to deal supply suggests continued selectivity for buyers navigating increasingly complex transactions.
- Equities
ai hardware stack follow the bottlenecks not the chips
Polaris argues the real AI investment value lies not in chip winners but one layer down—in supply-chain bottlenecks like advanced packaging, substrates, memory and power infrastructure. These capacity-constrained enablers, often mid/small-cap and globally dispersed, hold durable pricing power the fund is positioned to capture.
- Private Markets
Grid capex in the age of renewables and AI
Electricity grids have become the bottleneck for both the energy transition and AI-driven demand growth, requiring a forecast $5.9 trillion in global grid investment over 2026–2035. The opportunity spans transmission, distribution networks, and specialist grid services across developed and emerging markets.
- Fixed Income
A golden age for infrastructure debt?
Macquarie's Tom van Rijsewijk and Harlan Cherniak argue infrastructure debt is evolving from a niche into a core allocation, offering resilient, inflation-linked cash flows and higher yields amid bank retrenchment. Structural tailwinds from digitalisation, decarbonisation, deglobalisation and demographics underpin trillions in opportunities.
- Private Markets
What AI can't build: The enduring role of infrastructure in a disrupted world
Macquarie argues infrastructure's physical, capital-intensive, long-lived nature makes it resilient to AI substitution while benefiting from AI-driven demand for data centres, power grids and renewables. Even in a downside AI scenario, data centre power demand still grows, though risks include capital concentration and speculative pre-lease funding.
- Macro
NI + AI ‘Contemplating…stand by’
Macquarie argues two accelerating themes—the shift toward national interests (NI) and the AI build-out—support robust nominal growth but raise inflation risk, sharpened by the Strait of Hormuz closure. It advocates lower duration near-term, buying rate and credit dips, while warning of stretched AI assumptions and unprecedented index concentration.