Investing in the Infrastructure

Behind the AI Boom

Everyone wants exposure to artificial intelligence. Far fewer investors stop to ask what AI actually runs on.

Behind every chatbot response and every AI-generated image sits a physical supply chain: chips, memory, networking gear, buildings, cooling systems, and enough electricity to power a small country. AI is not just software; it is one of the biggest physical infrastructure build-outs in history, and it’s happening in real time.

During the California gold rush, the people who reliably made money weren’t always the prospectors; many were the merchants selling the picks, shovels, and denim jeans. Investors looking at the AI theme can apply the same lens. Rather than betting solely on who builds the smartest model, you can look at who supplies the infrastructure that makes AI possible in the first place.

The AI stack has six layers

Think of AI infrastructure as a stack, running from the silicon at the very bottom to the software furthest from it at the top. Each layer has its own companies, its own risks, and its own return drivers. Here’s how it breaks down.

1. Computer, Chips & Accelerators

This is the foundation of the entire AI economy: the specialised chips that actually do the mathematical heavy lifting of training and running AI models. GPUs, ASICs, TPUs and AI accelerators, plus the chip design and lithography needed to manufacture them.

Exposure type	Companies
Global	Nvidia, AMD, Broadcom, Marvell, Arm, TSMC, ASML, Alphabet, Amazon, Microsoft
ASX	BrainChip – BRN.ASX Brainchip Holdings Limited (BRN) is the first-to-market neuromorphic processor, Akida, mimics the human brain to analyse only essential sensor inputs at the point of acquisition, processing data with unparalleled efficiency, precision, and economy of energy. Weebit Nano – WBT.ASX Weebit Nano Ltd. (WBT) is a developer of semiconductor memory technology Its Resistive RAM (ReRAM) technology addresses the need for performance and lower power non-volatile memory (NVM) solutions in a range of new electronic products and devices being used in AI, Internet of Things (IoT) and wearable devices, automotive, industrial automation, robotics, neuromorphic computing, and many others.
ETFs	SEMI – SEMI.ASX Semiconductor ETF (SEMI) seeks to invest in companies that stand to potentially benefit from the broader adoption of tech-enabled devices that require semiconductors. This includes the development and manufacturing of semiconductors. ASIA – ASIA.ASX ASIA aims to track the performance of an index (before fees and expenses) comprising the 50 largest technology and online retail stocks in Asia (ex-Japan), including technology giants such as Alibaba, Tencent, Baidu and JD.com. NDQ – NDQ.ASX NDQ aims to track the performance of the Nasdaq 100 Index (before fees and expenses). The Nasdaq 100 comprises 100 of the largest non-financial companies listed on the Nasdaq market, and includes many companies that are at the forefront of the new economy. FANG – FANG.ASX The Global X FANG+ ETF (FANG) seeks to invest in companies at the leading edge of next-generation technology that includes household names and newcomers. GXAI – GXAI.ASX The Global X Artificial Intelligence ETF (GXAI) seeks to invest in companies that potentially stand to benefit from the further development and utilisation of artificial intelligence (AI) technology in their products and services, as well as in companies that provide hardware facilitating the use of AI for the analysis of big data.

This layer gets the most media attention, and carries some of the richest valuations. It’s also the layer most exposed to the “silicon problem”: GPUs are punishing, high-heat assets with a useful life of roughly one to three years before they’re physically worn out or technologically outpaced.

2. Memory & Interconnect

AI chips are only as fast as the data that can be fed to them. This layer covers the high-bandwidth memory, optical interconnects, networking chips, fibre and switches that move enormous volumes of data between chips, servers and data centres.

Exposure type	Companies
Global	Micron, SK Hynix, Samsung, Astera Labs, Credo, Arista, Ciena, Coherent, Corning
ASX	Megaport – MP1.ASX Megaport Limited (MP1) is a Company engaged in provisioning of on-demand interconnection services, provisioning of internet exchange services addition and integration of new service providers into the Ecosystem, development of product features and API integration with key partners and continuing to expand the geographic footprint of its Network and Marketplace. Weebit Nano – WBT.ASX Weebit Nano Ltd. (WBT) is a developer of semiconductor memory technology Its Resistive RAM (ReRAM) technology addresses the need for performance and lower power non-volatile memory (NVM) solutions in a range of new electronic products and devices being used in AI, Internet of Things (IoT) and wearable devices, automotive, industrial automation, robotics, neuromorphic computing, and many others.
ETFs	SEMI – SEMI.ASX Semiconductor ETF (SEMI) seeks to invest in companies that stand to potentially benefit from the broader adoption of tech-enabled devices that require semiconductors. This includes the development and manufacturing of semiconductors. ASIA – ASIA.ASX ASIA aims to track the performance of an index (before fees and expenses) comprising the 50 largest technology and online retail stocks in Asia (ex-Japan), including technology giants such as Alibaba, Tencent, Baidu and JD.com.

3. Data Centres

Every AI model needs somewhere to physically live. This layer is the real estate and physical shell of AI: land, powered sites, buildings, racks, servers, cloud hosting and connectivity.

Exposure type	Companies
Global	Equinix, Digital Realty, CoreWeave, Nebius
ASX	NextDC – NXT.ASX NEXTDC Limited (NXT) engages in the development and operation of independent data centres in Australia and the Asia-Pacific region. Goodman Group – GMG.ASX Goodman Group (GMG) is a Company engaged in investment in industrial property (either directly or in partnerships with other investors) and management services provided to the partnerships (including investment management, property management and development management). DigiCo REIT – DGT.ASX Digico Infrastructure Reit (DGT) is a a diversified owner, operator and developer of data centres, with a global portfolio and broad investment mandate across Stabilised, Value-add and Development opportunities. Macquarie Technology – MAQ.ASX Macquarie Technology Group Limited (MAQ) is engaged in the provision of telecommunication, cloud computing, cybersecurity and data center services to corporate and government customers within Australia. Megaport – MP1.ASX Megaport Limited (MP1) is a Company engaged in provisioning of on-demand interconnection services, provisioning of internet exchange services addition and integration of new service providers into the Ecosystem, development of product features and API integration with key partners and continuing to expand the geographic footprint of its Network and Marketplace. Dicker Data – DDR.ASX Dicker Data Limited (DDR) is an Australian-owned and operated technology hardware, software, and cloud distributor. It is engaged in distribution of IT hardware, software, cloud, access control, surveillance and emerging technology solutions for the corporate and commercial market.
ETFs	ATEC – ATEC.ASX ATEC aims to track the performance of the S&P/ASX All Technology Index (before fees and expenses). The Index provides exposure to leading ASX-listed companies in a range of tech-related market segments such as information technology, consumer electronics, online retail and medical technology. AINF – AINF.ASX The Global X Artificial Intelligence Infrastructure ETF (AINF) offers targeted exposure to the physical and operational backbone enabling AI’s global expansion. While most AI investments focus on chips or platforms, AINF looks underneath the surface at the energy, data, and materials infrastructure powering this transformation.

Unlike chips, data centre buildings, cooling systems and grid connections have a much longer useful life , often 15 to 20 years. That makes this layer closer in character to the “durable infrastructure” that has survived every past technology boom, from railway tracks to dark fibre.

4. Cooling & Power Management

AI chips generate enormous amounts of heat, and keeping a data centre running requires serious thermal and electrical engineering: liquid cooling, thermal management, UPS systems, switchgear and electrical infrastructure.

Exposure type	Companies
Global	Vertiv, Eaton, Schneider Electric, Modine, nVent, Dell, HPE, Super Micro
ASX	No obvious pure-play
ETFs	AINF – AINF.ASX The Global X Artificial Intelligence Infrastructure ETF (AINF) offers targeted exposure to the physical and operational backbone enabling AI’s global expansion. While most AI investments focus on chips or platforms, AINF looks underneath the surface at the energy, data, and materials infrastructure powering this transformation.

Notably, there is currently no clean ASX pure-play in this layer , for direct exposure, Australian investors are largely looking offshore or via diversified ETFs.

5. Electricity, Grid & Infastructure

AI’s biggest bottleneck may not be chips at all , it’s power. Data centres are enormous, round-the-clock electricity consumers, and this layer covers generation, transmission, transformers, batteries, nuclear, small modular reactors (SMRs) and uranium.

Exposure type	Companies
Global	GE Vernova, Constellation, Vistra, Talen, NextEra, Quanta, Bloom Energy, Cameco, Oklo, NuScale
ASX	Paladin Energy – PDN.ASX Paladin Energy Limited (PDN) is a uranium production and exploration company with projects in Australia, Canada, and Africa, including the Langer Heinrich Mine in Namibia. Boss Energy – BOE.ASX Boss Energy Ltd (BOE), is a multi-mine uranium producer, with 100% owned Honeymoon project in SA and 30% ownership in Alta Mesa project in South Texas. AGL – AGL.ASX AGL Energy Limited (AGL) supplies around 4.6m energy, telecommunications and Netflix customer services. AGL operates Australia’s private electricity generation portfolio within the National Electricity Market, comprising coal and gas-fired generation, renewable energy sources such as wind, hydro and solar, and batteries and other firming and storage technology. Origin – ORG.ASX Origin Energy Limited (ORG) is involved in the operation of energy businesses including exploration and production of natural gas, electricity generation, wholesale and retail sale of electricity and gas, and sale of liquefied natural gas. GenusPlus – GNP.ASX GenusPlus Group (GNP) provides power and communications infrastructure services across Australia. The organisation undertakes the design, construction and maintenance of electrical transmission networks, distribution networks, substations and battery systems. APA – APA.ASX APA Group (APA) is a Company that operates a portfolio of gas, electricity, solar and wind assets. Its principal activities include Energy Infrastructure, Asset Management and Energy Investment.
ETFs	AINF – AINF.ASX The Global X Artificial Intelligence Infrastructure ETF (AINF) offers targeted exposure to the physical and operational backbone enabling AI’s global expansion. While most AI investments focus on chips or platforms, AINF looks underneath the surface at the energy, data, and materials infrastructure powering this transformation. URNM – URNM.ASX URNM aims to track the performance of an index (before fees and expenses) that provides exposure to a portfolio of leading companies in the global uranium industry. ATOM – ATOM.ASX The Global X Uranium ETF (ATOM) offers investors access to a broad range of companies involved in uranium mining and the production of nuclear components, including those in extraction, refining, exploration, or manufacturing of equipment for the uranium and nuclear industries. ACDC – ACDC.ASX The Global X Battery Tech & Lithium ETF (ACDC) offers investors exposure to global companies developing electro-chemical storage technology and mining companies producing battery-grade lithium.

This layer has become one of the most closely watched in the entire AI theme, as hyperscalers increasingly sign long-term power deals , including nuclear , just to secure enough electricity to keep expanding.

6. Broad AI & Technology Exposure

Finally, there’s the layer most investors already know: the cloud platforms, AI software and hyperscalers actually building and deploying AI models on top of all this infrastructure.

Exposure type	Companies
Global	Microsoft, Alphabet, Amazon, Meta, Apple, Nvidia
ASX	Limited direct exposure
ETFs	NDQ – NDQ.ASX NDQ aims to track the performance of the Nasdaq 100 Index (before fees and expenses). The Nasdaq 100 comprises 100 of the largest non-financial companies listed on the Nasdaq market, and includes many companies that are at the forefront of the new economy. FANG – FANG.ASX The Global X FANG+ ETF (FANG) seeks to invest in companies at the leading edge of next-generation technology that includes household names and newcomers. GXAI – GXAI.ASX The Global X Artificial Intelligence ETF (GXAI) seeks to invest in companies that potentially stand to benefit from the further development and utilisation of artificial intelligence (AI) technology in their products and services, as well as in companies that provide hardware facilitating the use of AI for the analysis of big data. RBTZ – RBTZ.ASX RBTZ aims to track the performance of an index (before fees and expenses) that includes leading global companies involved in the production or use of robotics and robotics-focused AI products and services. ROBO – ROBO.ASX The Global X ROBO Global Robotics & Automation ETF (ROBO) seeks to invest in companies that potentially stand to benefit from increased adoption and utilisation of robotics and artificial intelligence, including those involved with industrial robotics and automation, non-industrial robots, and autonomous vehicles.

The Bottom Line

The AI investment story is really an infrastructure story. Different layers of the stack carry different risks and different return drivers, chips wear out fast, buildings and grid connections last decades, and software sits on top of it all.

A few things worth keeping in mind:

Watch for overlap. Many thematic AI ETFs hold significant, overlapping positions in the same handful of mega-cap names, particularly Nvidia, so “diversified” AI exposure can be less diversified than it looks.
Know which layer you actually own. Buying “AI” broadly can mean very different things depending on whether you’re exposed to chips, data centres, power, or software.

Lifespan matters. As we’ve explored in our article on AI and the great technology cycle, the silicon at the bottom of the stack has a far shorter useful life than the buildings and power infrastructure around it , which has implications for where long-term value may ultimately settle.

What you learn here has been used in our Trade for Good software.
Click on the button to find our software education videos.

Software Videos

You can read more of our educational articles in the Trade for Good Learn section
Trade for Good Learn