In Australian investing circles, there’s a quiet discussion going on that doesn’t make headlines around the world but lingers at industry breakfasts in Sydney and Melbourne. There is no Nvidia available in the local market. There is no Palantir on it. What it does have, more and more, is a small but remarkably tangible group of AI-related businesses that own the unglamorous infrastructure required by the entire sector. As you observe this from the southern hemisphere, you begin to notice an oddity. Language models and chatbots are not the foundation of the ASX’s AI trading. Data centers, distribution routes, network ports, memory chips, and forensic software form its foundation. More plumbing, less spectacle.
At the top of that list, NEXTDC has a market capitalization of approximately AU$8.07 billion. The operator, which has its headquarters in Sydney, operates sixteen operational data centers and plans to open more in Adelaide, Darwin, Port Melbourne, Kuala Lumpur, and Auckland. The December memorandum of understanding with OpenAI, a roughly AU$7 billion sovereign AI hyperscale campus and GPU supercluster planned for the company’s S7 site in Sydney, was the detail that ought to have generated more buzz than it did. That is a significant footnote. Having OpenAI publicly commit to NEXTDC’s site is closer to validation than anything the ASX has produced in years, especially for a nation still debating whether it has the energy grid to support large-scale AI workloads.
Next is Dicker Data, which sounds more like a suburban IT store from the 1990s than the AI distribution hub it has subtly transformed into. With a market capitalization of AU$1.56 billion, the company generated AU$3.9 billion in gross revenue last year, an increase of about 15%. This is partially due to the fact that Dicker Data has spent two decades establishing the channel relationships necessary to deliver enterprise-grade AI deployments, which resellers throughout Australia are rushing to deliver. Its supply lines are used by Cisco, Dell, and ResetData’s AI-F1 sovereign factory. There is a perception that the company’s lack of sexiness is the reason it has been undervalued. Seldom is distribution.
Megaport’s situation is distinct, bordering on philosophical. The network-as-a-service provider, which has its headquarters in Queensland and operates in 26 countries with over a thousand enabled locations, has a customer list that resembles a list of every significant cloud and hyperscaler. It has recently caused difficulties for investors. In the first half of fiscal 2026, the company reported strong recurring revenue growth; however, a net loss of AU$19.1 million shows how much it is spending to position itself for AI workloads, including a US$70 million acquisition of a fast-deploy compute provider. It’s the type of stock that divides rooms.
It’s important to state clearly that Weebit Nano is the speculative one. The dual-listed Israeli-Australian business, valued at AU$811.97 million, does not produce artificial intelligence products. It produces ReRAM, a non-volatile memory technology intended for neuromorphic computing and edge AI applications. The GlobalFoundries 22FDX integration progress, the licencing agreements with Onsemi and Texas Instruments, and the AU$80 million capital raise earlier this year all indicate that the company is attempting to commercialize a genuine piece of physics rather than ride a wave. It’s really unclear if ReRAM becomes the norm or becomes a footnote. However, the cap table is patient and the wager is intriguing.
With investigative analytics software utilized by tax authorities, regulators, and law enforcement agencies, NUIX completes the list at AU$406.73 million. The company recently signed a multi-year agreement with the German state of Rhineland-Palatinate and acquired Linkurious using graph-AI. Compared to most enterprise software peers, forensic tools are less consumer-facing, but once they are integrated into a legal workflow, they tend to remain there.
Local access to international capital flows without the US’s built-in valuation premium is an undervalued advantage for Australian investors. The field is subtly tilted by sovereign AI demand, an abundance of renewable energy in places like Tasmania (where Firmus is reportedly preparing an IPO valued near AU$6 billion), and a regulatory environment less susceptible to abrupt export controls. The businesses aren’t inexpensive, and the majority of them don’t make as much money as conservative Australian investors would like. However, the deals being signed in 2026 indicate that the local AI trade has progressed beyond the slogan stage, and the cluster and customers are real. Most of the unanswered questions should be addressed in the next two years.
