As businesses and industries around Australia are increasingly turning to AI for solutions, Nvidia is bracing itself for a significant growth in data centre computing power in Australia. The tech firm will partner with a number of Australian cloud and data centre providers to boost the nation’s capacity to process AI workloads, and the technology could bring national AI computing power to a level far exceeding what is currently possible.
The planned development showcases how data centres are increasingly becoming a vital part of Australia’s technology ecosystem. The introduction of generative AI, advanced analytics, machine learning and other computation-heavy applications has significantly boosted the need for high-performance and high-reliability computing facilities as companies integrate these technologies. Nvidia’s latest moves suggest Australia is a key market for the next stage of building AI infrastructure.
The proposed expansion is in partnership with Australian firms such as Firmus, CDC, NEXTDC and AirTrunk. The two companies aim to reach up to 2GW of computing power related to AI by 2027. The planned buildout would be a significant expansion in the current capacity of data centers in Australia.
Australia currently has approximately 1.6 gigawatts of computing capacity, according to industry estimates. The proposed 2GW of additional capacity for AI therefore is a massive expansion to the country’s infrastructure. The figure also shows the rapid evolution of demand for computing power as AI takes a deep-rooted place in business operations.
The extra infrastructure would support Australia’s ever-expanding need for AI computing, Nvidia said. The company also stated that the data centres being built by the partners would be running its own data centre technology platform called DSX, which is an architecture developed to enable the deployment and management of data centres focused on AI.
As more industries adopt AI, so does the need for high-performance computing, Nvidia said in a statement.
The expansion is at a time when Australia is trying to position itself as a hub for investment in AI and other technologies. Data centres are a crucial component of that vision as the computing power required for AI systems, especially when scaling up, is significant. Local infrastructure can also enable companies to make use of computing services more effectively, and enhance the country’s overall digital economy.
The Australian rollout is part of a wider Nvidia strategy for the expansion of the AI infrastructure market. Nvidia is recognized for its cutting-edge graphics processing units (GPUs), which are crucial for Artificial Intelligence (AI) systems. The processors are essential to the infrastructure of generative AI and other new technologies, capable of making the more complex calculations needed to train and run sophisticated AI models.
The company’s increasing presence in data centre development is another example of the evolution of the traditional tech supply chain in the age of AI. Demand isn’t just for processors and servers. Companies require full-scale infrastructure ecosystems, with high-performance data centres, networking, cooling, electricity power and cloud computing.
With Australia’s geography and energy resources offering potential for additional data centre growth, there are also pragmatic issues to consider. Power usage is significant at large computing facilities, especially if they are used to run heavy AI workloads. Depending on the technology and location, cooling these facilities can also be water and energy intensive.
These issues have come to the fore as governments and communities discuss the pace of data centre infrastructure growth. New facilities can stimulate investment, offer jobs and enhance the digital infrastructure, but at the same time can impose new strains on electricity networks and local resources. Discussions about Australia’s ambitions for AI are likely to continue to feature questions around energy consumption, water use, land requirements and environmental impact.
This is especially important because the amount of AI computing power needs can rapidly increase. AI applications can place significant demands on computing resources and are already a strain on large data centres that are common in traditional digital services. Running a large model and training it for millions of users can consume a lot of processing power, leading to the need for a significant amount of electricity and dedicated cooling.
The potential expansion also underscores the need for global tech giants to partner with local infrastructure providers. Instead, Nvidia will build each facility with the help of existing Australian data centre operators who have built and operated data centres before. Such collaborations can facilitate the scaling of infrastructure and benefit from local expertise and existing infrastructure and networks in the geographical area.
In Australia, improved local computing power could help drive the adoption of cutting-edge AI tools.Increased domestic computing power may increase the reach of these sophisticated AI tools for Australian businesses. From financial services to healthcare, manufacturing to telecom and professional services, industries are increasingly trying out AI tools. More compute infrastructure can enable this adoption and offer enterprises a range of possibilities to run more intensive workloads.
In addition, there are broader economic impacts if Australia can establish a viable AI infrastructure ecosystem. Cloud Computing can service data centres, software development, research and other technology services. A bigger supply base would help support companies to open more AI-related businesses in Australia, which could help bolster Australia’s presence in the broader APAC technology market.
