From Data Centres to Patents: Why Australia and South East Asia are key patent filing jurisdictions for AI innovations

Fixed infrastructure and a finite set of jurisdictions

For many software inventions, it has historically been difficult to confidently identify where the invention is likely to be carried out. In cloud environments, processing may be distributed dynamically across multiple regions, and the physical location of back end compute could be anywhere. From a patent filing perspective, that has made it difficult to predict where an invention would ultimately be performed, or where competitors may attempt to implement particular steps to avoid infringing existing patents.

However, the training and large scale inference demands of AI systems have changed that. They require specialised hardware, substantial power supply, and purpose built facilities that can provide dense GPU clusters, high bandwidth networking, and sustained energy and cooling capacity that far exceed traditional computing environments. As a result, AI workloads tend to be concentrated in only those jurisdictions capable of supporting this infrastructure.

In the Asia Pacific region, those jurisdictions are becoming well defined. Australia is investing heavily in hyperscale AI infrastructure[AH2] , supported by major capital commitments from global providers. And South East Asian jurisdictions such as Singapore, Malaysia, Indonesia and Thailand are also expanding rapidly as alternative capacity hubs.

For many AI enabled inventions, this means that there is now a more readily identifiable set of jurisdictions in which the invention is likely to be performed at scale.

This means that IP professionals can use this information to assess where an AI invention would realistically be deployed, given its compute requirements, and to guide decisions on where national patent applications should be filed to better align protection with deployment of AI inventions.

Filing decisions can therefore be anchored to a finite set of infrastructure capable jurisdictions, rather than being spread across an indeterminate global footprint or limited by uncertainty as to where the invention is performed. This, in turn, elevates the importance of those jurisdictions, irrespective of whether they are the largest end user markets.

Filing to control the compute layer

If core steps of an invention depend on AI grade infrastructure, then filing in major data centre jurisdictions can have significant practical impact. Protection is no longer limited to covering end user markets, but extends to capturing the locations where a competitor will want (and in many cases need) to deploy systems at scale.

Accordingly, where an AI invention depends on data centre level processing, securing protection in these jurisdictions provides a form of practical control over both training and deployment activities.

For example, filing in these jurisdictions means:

• IP protection can be secured in where the invention must be executed to achieve commercial scale
• protection can be aligned with the location where technically significant steps are performed
• enforcement against infrastructure based deployment can be supported

In effect, protection at the infrastructure layer can be highly consequential. If a competitor cannot lawfully operate their system in jurisdictions that host the necessary compute, their ability to deploy commercially meaningful solutions at scale is limited.

The same logic applies once a model is deployed. Whether an AI system is exposed through an API, embedded in software, or delivered as a service, it continues to run on underlying data centre infrastructure. Users may be global, but the processing of requests and generation of outputs still occur where the compute is physically located. Each inference request is ultimately executed on hardware within those facilities.

As a result, commercially meaningful use of the invention remains tied to a relatively small number of jurisdictions. Even where users are distributed across multiple countries, competitors may still need to operate within these infrastructure locations. Securing protection in those jurisdictions therefore captures both the back end processing and the ongoing operation of the AI system as it is made available to users, while also strengthening the patentee’s ability to limit competitive market entry by restricting where competing systems can realistically operate at scale.

Australia and South East Asia as priority filing locations

Accordingly, jurisdictions that host critical data centre infrastructure are emerging as priority filing locations. Filing in these locations is not just about capturing local activity, but about positioning in the environments where competitors will need to deploy and operate systems at scale.

As discussed in our recent article From 'nice to have' to 'must file': IP and Australia's AI infrastructure boom, this is particularly evident in Australia. But the South East Asia region is also becoming relevant, with data centre capacity rapidly expanding, reinforcing the strategic importance of these jurisdictions.

The South East region already hosts a substantial and growing base of infrastructure, with continued investment driven by AI and broader digital demand. Currently, Singapore has South East Asia’s largest capacity of data centres, while Malaysia and Thailand are quickly emerging as viable alternatives, supported by available land, power, and government incentives.¹

Collectively, these jurisdictions form a practical infrastructure layer for regional AI deployment and accordingly, a relatively predictable and commercially meaningful set of filing destinations. Jurisdiction such as Australia, Singapore, Malaysia, and Thailand can should therefore be seen less as standalone markets, and more as access points to essential compute capability and central to an effective AI patent strategy.

Coordinated portfolio management across key AI jurisdictions

Putting this into practice requires coordination across jurisdictions with different requirements and timelines. This is a skill FB Rice has honed over the years through our South East Asia hub offering.

FB Rice acts as a central point of contact to deliver the highest quality service, and consistency in prosecution strategy across the entire Australian and South East Asia portfolio, streamlining the workflow and delivering a more seamless, efficient experience for clients across the region.

The expansion of AI infrastructure across Australia and South East Asia point to a shift in how AI related patent protection should be approached. Rather than spreading filings across broad and uncertain geographic footprints, IP practitioners should focus on jurisdictions that host the infrastructure required to support large scale AI systems. This approach not only aligns protection with where technically and commercially significant activity occurs, but also offers a means of control over how and where competitors can operate.

As investment in AI infrastructure continues, particularly across Australia and South East Asia, early and coordinated filing in these jurisdictions will become an important component of effective and forward looking patent strategy.

If you would like to explore how to position your AI inventions in key data centre jurisdictions, or obtain tailored guidance on filing across Australia and South East Asia, please contact us.