Owning the algorithm: Why Australia must build its own AI capabilityBy David Hall-Johnston, chief technology officer Dell Technologies ANZ.Advertiser content for Dell Technologies.This year is shaping up to be a defining moment for where Australia’s AI systems are built, hosted and governed. Industry is scaling domestic AI capacity, as governments move to define national AI priorities. This means now is the time to move beyond AI hosted on someone else’s servers, trained on someone else’s data.AI-related patents in Australia grew from 170 in 2015 to roughly 629 in 2024, according to the CSIRO. Sounds impressive, but intense competition means that’s about 0.2 per cent of the global total. At this stage, we are more AI takers than makers.This is the year the question of Australian control over AI infrastructure moves beyond niche concern. Countries around the world, including Australia, are anchoring their AI capabilities in domestic infrastructure.Control is now a strategic necessity.Last November’s launch of the peak body Data Centres Australia signals an industry ready to scale local infrastructure. The federal government’s National AI Plan, announced in December, also provides more certainty. As does its newly released data centre expectations that encourages organisations to act in Australia’s interest, support clean energy, use resources responsibly, create local jobs and boost innovation.Owning the intelligence layer is criticalThe fact is, building Australian-controlled AI capability belongs in the same conversation as Future Made in Australia, the framework seeking to secure local energy and manufacturing capability. In a similar vein, we must build the digital foundations for an intelligence economy. This requires homegrown data, domestic infrastructure and Australian expertise.While often discussed through a government lens, sovereign AI is about more than public services. The same infrastructure accelerates enterprise capability across the economy, from energy and manufacturing to finance and healthcare.We do not need to build every chip or cloud from scratch, but owning the intelligence layer is key. Data, algorithms and governance systems will determine how AI shapes Australia. If we fail to take ownership, we risk dependency on offshore platforms. This means increased exposure to bias and opacity in foreign-trained models, issues with data sovereignty and security, and the loss of intellectual property and competitive advantage.A new wave of capability made for AustraliaAs AI moves from experimentation to scaled deployment, attention is shifting to AI factories: resilient environments where models are built, governed and operated at scale.Australia needs more organisations like Maincode. The Melbourne-based start-up recently announced MC-2, a sovereign AI infrastructure platform. It uses Dell Technologies servers, along with other technology partners including AMD and Telstra InfraCo. to provide the compute, storage and orchestration needed to build and run large language models locally.This will be critical as agentic AI matures. Large foundation models provide a base, but organisations need smaller, fine-tuned models operating safely within Australian-controlled environments to achieve specific outcomes.MC-2 is not a generic cloud platform. It is a controlled environment where organisations can train, tune, deploy and scale secure models using Australian data. It demonstrates how global technology partners and local innovators can create practical, scalable and commercially viable Australian-made AI capability. For high-risk applications, from autonomous systems to critical infrastructure, hosting AI outside national borders is not the answer.Maincode is also developing Matilda, a home-grown large language model that is being trained within Australia. It will better reflect local data, regulation and linguistic context. Its purpose is not to recreate what is happening overseas. Instead, it gives organisations a platform to build, customise and deploy their own language models securely in Australia. Designed for comprehension rather than conversation, it focuses on understanding, reasoning and domain accuracy rather than broad general-purpose chat.The economic case for Australian-made AILocal AI environments will create trusted zones for collaboration, allowing organisations – and even partners across borders – to work together in controlled settings where data and AI agents can interact securely.As AI adoption accelerates, Australian hosted infrastructure is likely to underpin entire ecosystems of innovation. Its economic impact is expected to be significantly larger than many anticipate today, as every advanced AI use case ultimately depends on trusted infrastructure.AI infrastructure, like MC-2 and Australian-made models like Matilda, will create new IP, support highly skilled jobs and speed up the move from research to industry application. These capabilities will be important in sectors such as healthcare, energy, government and finance where compliance, trust and the proper handling of sensitive data all require domestic control.Shifting from consuming global AI to producing our own intelligence systems will require coordinated effort from government and industry, along with access to global tools and local expertise. Australia will need to invest in AI ecosystems, not just AI adoption, if it wants to secure the full economic and social benefits of AI.Building AI capability made for Australia and its local economic and cultural needs is key to securing our digital future. The next generation of competitiveness will come from who owns the algorithms that shape the economy. With projects like MC-2 and Matilda, Australia is showing that it can manufacture intelligence, not just import it.Dell Technologies (NYSE: DELL) helps organisations and individuals build their digital future and transform how they work, live and play. The company provides customers with the industry’s broadest and most innovative technology and services portfolio for the AI era.