AI Investment Across Asia Pacific: The Spending Surge and How Australian Investors Can Access It
Investment in artificial intelligence across Asia Pacific reached record levels through 2025 and 2026, spreading well beyond the United States and China into Singapore, Australia, New Zealand, and the wider region. Direct data centre investment across Asia Pacific hit a record US$11.6 billion in 2025, according to CBRE's 2026 Asia Pacific Data Centre Trends report, and technology firms operating in the region plan to lift AI-related capital spending by 61% in 2026. The most recent marker landed in September 2026, when Databricks committed more than US$350 million to Singapore over three years.
For wholesale and sophisticated investors in Australia and New Zealand, that build raises a practical question. When the companies driving it are either public mega-caps too large for the spending to move their valuations, or private companies the public markets cannot reach, where does a pure position in APAC AI growth actually come from? The answer maps to where the money is going, so this piece walks the region market by market, then sets out the routes an Australian or New Zealand investor has to participate.
Why is AI investment growing across Asia Pacific?
The spending is being driven by forces that look durable rather than cyclical.
Start with government policy. Singapore's National AI Strategy, Australia's National AI Plan, and comparable programmes across the region have turned AI capacity into a matter of national economic policy. Governments are funding domestic capability and, in several markets, requiring that data stay onshore, which forces global providers to build locally rather than serve the region from hubs elsewhere.
Enterprise adoption is the second force, and it has shifted from pilots into production. Forrester forecasts that Asia Pacific technology spending will grow 9.3% in 2026, with the region spending more than US$437 billion on new technology between 2026 and 2030, and computer equipment the fastest-growing category at 13.7% as organisations build out AI-ready infrastructure. Companies across financial services, telecommunications, and the public sector have stopped experimenting and started deploying, and deployment at scale costs far more than experimentation did.
Underneath both sits the weight of capital committing to the region. McKinsey's research on data centre demand has Asia Pacific accounting for roughly 30% of global data centre capacity expansion over the next five years. Alibaba alone committed at least US$52 billion to cloud and AI infrastructure globally over three years, and Chinese and Western hyperscalers are now expanding across Southeast Asian markets in parallel. The region has shifted from a place these companies served to a place they build in.
The Databricks signal: the value sits above the data centre
Databricks' Singapore commitment is worth reading closely, because it points to where the durable value in this build actually accrues.
The company said in September 2026 that it would invest more than US$350 million in Singapore over three years, quadruple its local office with a new 32,000-square-foot regional headquarters, and roughly double its Singapore workforce from 250 to more than 500 people, according to its announcement. Singapore serves as its regional hub for Asia Pacific and Japan. New regional customers named alongside the investment include Singtel, Singapore Customs, and iFAST Corporation, joining existing users such as Standard Chartered, GovTech Singapore, LG Electronics, and Toyota.
Most coverage of the APAC AI build focuses on the physical layer, the gigawatts of data centre capacity and the hyperscalers pouring concrete. That layer matters, but it is largely a landlord business, and much of it is already owned by the largest public companies in the world. The Databricks investment is a marker for a different layer sitting on top: the enterprise data and AI software that companies actually pay recurring fees to run. Databricks builds the tools that let an organisation put its own data to work with AI under proper governance, and that governance layer is precisely what regional enterprises now need as they move into production.
Here is the observation that follows, and one I have not seen made cleanly in the Australian or New Zealand context. The APAC AI build is usually told as an infrastructure story, and infrastructure is where public-market investors can already participate by owning the hyperscalers and the data centre operators. The concentrated, pure-play exposure to the enterprise-AI layer, the software capturing the recurring spend, sits disproportionately in private companies. Databricks is the clearest example. It is one of the largest software companies in the world yet to list, most recently valued at around US$130 billion, generating positive free cash flow, and it is the company writing the US$350 million cheque into Singapore. An investor in Sydney or Auckland reading about this build cannot buy Databricks on the ASX or the NZX. That gap between where the value accrues and where public investors can reach is the whole point of the section below.
What is driving AI investment in Australia and New Zealand?
Australia and New Zealand are not spectators to the regional build. Both have drawn some of the largest technology infrastructure commitments in their history over the past 18 months.
In Australia, Microsoft committed A$25 billion (about US$18 billion) through to the end of 2029 to expand its Azure AI and cloud capacity, its largest investment in the country on record, announced during Satya Nadella's April 2026 visit to Sydney. The commitment expands Microsoft's Australian data centre footprint by more than 140% and includes a pledge to equip three million Australians with AI skills by 2028. Amazon Web Services had already pledged A$20 billion to Australian data centre capacity over 2025 to 2029, and OpenAI announced a multi-billion-dollar Australian initiative in December 2025. By mid-2026, industry reporting placed Australia among the world's largest destinations for data centre investment behind only the United States, and Westpac economists put the country's total data centre investment pipeline above A$155 billion, a scale that rivals the mining investment boom.
New Zealand's build is smaller in absolute terms but significant for its size. Amazon opened its AWS Asia Pacific (New Zealand) Region backed by an investment of more than NZ$7.5 billion (about US$4.4 billion), which Amazon estimates will add around NZ$10.8 billion to New Zealand's GDP and support more than 1,000 full-time jobs a year. Local firms such as Xero and Kiwibank gain an onshore option to run and store data, which lowers the barrier for New Zealand organisations to adopt AI services that previously ran offshore.
The demand side across both markets is already live. Research from Salesforce puts AI agent deployment at roughly half of organisations across Australia and New Zealand, often with limited formal governance in place. That combination, heavy adoption with thin governance, is exactly the problem the enterprise-AI software layer is built to solve, which is why the companies selling data and AI governance tools are expanding into the region rather than waiting for it to mature.
How can Australian and New Zealand investors get exposure?
An investor convinced by the trend has several routes, and they differ in how directly they track the theme.
The most accessible is public equity in the hyperscalers and chipmakers. Buying Microsoft, Amazon, Alphabet, or Nvidia gives exposure to companies spending heavily on the APAC build. The limitation is dilution of the theme. Microsoft's A$25 billion Australian commitment is significant in absolute terms and a rounding error against its multi-trillion-dollar market capitalisation, so the AI-in-APAC story barely registers in the share price.
Listed infrastructure offers a closer read. Data centre operators and real estate investment trusts with regional exposure, along with a growing set of AI-themed exchange-traded funds, track the physical build more directly than a mega-cap does. This is the landlord layer, and it is a legitimate way to own the concrete and power underpinning the region's AI capacity.
Managed venture and private equity funds give pooled, professionally selected exposure to private companies across the theme, in exchange for long lock-up periods and layered fees.
The route that reaches the private software layer directly is secondary access to specific late-stage companies. Platforms operating under an Australian Financial Services Licence source shares in established private companies from existing holders and structure them into single-company vehicles for wholesale and sophisticated investors. NonPublic sits in this group, and the companies available through it include Databricks, the same company anchoring the Singapore investment described above, alongside names such as OpenAI, Moonshot and Mercor. For an investor who wants a concentrated position in a specific driver of the APAC build rather than diversified index exposure, a secondary vehicle is the most direct route on offer.
Access carries a clear limit worth stating plainly. Secondary supply exists only where existing shareholders are willing to sell, so the newest or most tightly held private rounds are often unavailable at any price. Where these vehicles work is with larger, established private companies that have a functioning secondary market and a visible path toward an eventual listing. Databricks, with its scale, positive cash flow, and frequently reported IPO speculation, is the kind of name that fits.
The risks worth weighing
Private market investing in this theme carries the same structural risks that apply to the asset class generally, and the AI build does not suspend them. Positions are illiquid and typically held for several years before any liquidity event. Valuations for private companies can move sharply and are harder to verify than listed prices. Concentration in a single company or a single theme raises the stakes on getting the specific pick right.
The APAC angle adds two more considerations. Currency exposure is real for an Australian or New Zealand investor holding a US-dollar-denominated position in a US-headquartered company expanding into Asia. And the build itself faces a genuine constraint in power availability, which CBRE and others flag as the binding limit on how fast regional data centre capacity can actually come online. Strong structural demand does not guarantee that every company positioned to serve it will convert that demand into returns.
None of this argues against exposure. It argues for sizing positions against verifiable fundamentals, revenue, margins, market share, rather than against the momentum of the theme, and for treating each company as a separate decision rather than buying "APAC AI" as a single undifferentiated bet.
What it adds up to for Australian and New Zealand investors
The spread of AI investment across Asia Pacific is one of the clearest structural shifts in global technology spending, and the numbers behind it, US$11.6 billion in record data centre investment, a 61% rise in regional AI capex, more than US$437 billion in forecast technology spending through 2030, are large enough that the trend does not depend on any single company or any single year holding up.
The harder question is not whether the build is real but how to own it in a form that actually tracks it. Public markets offer the infrastructure landlords and the mega-caps whose AI spending is diluted across everything else they do. The enterprise-AI software layer, where a meaningful share of the recurring value accrues, remains largely private, and Databricks pouring US$350 million into Singapore is a reminder of exactly which tier of company is driving the regional build. For wholesale and sophisticated investors in Australia and New Zealand who want a concentrated position in that tier, the access exists through secondary vehicles, within the limits of what private supply allows.
Investors who want to understand which private companies are currently accessible, and on what terms, can create a free NonPublic account, or read the 2025 Private Markets Landscape report for a fuller view of how the asset class is structured. For those still weighing whether private markets suit their circumstances at all, the NonPublic platform overview sets out how access works and who qualifies.
NonPublic Pty Ltd (ABN 49 607 216 928) holds Australian Financial Services Licence #482668. Investments are available to wholesale and sophisticated investors as defined under the Corporations Act 2001. This content is general in nature and does not constitute financial product advice. It does not take into account your objectives, financial situation, or needs. Investing in private markets involves significant risk, including the potential loss of your entire investment. Past performance is not a reliable indicator of future results. You should obtain independent financial advice before making any investment decision.
