Australia is inadvertently ceding AI infrastructure leadership to China by maintaining copyright laws that tech companies are already circumventing overseas, while Chinese open-source models like DeepSeek capture roughly one-third of global AI use. The country faces a choice: reform copyright to attract data center investment, or watch AI development happen without it.

The core problem is straightforward. AI companies need massive amounts of creative content to train their models, but Australia's restrictive copyright framework makes them reluctant to build data centers locally. Instead, they train their models in other countries using the same Australian content anyway, leaving the country with no economic benefit, no job creation, and no influence over how AI develops.

Why Are Chinese AI Companies Moving So Fast?

While Australia debates copyright reform, Chinese AI developers are accelerating their open-source push. According to reporting cited in Source 1 from Bloomberg, The Wall Street Journal, and The New York Times, DeepSeek released preview versions of V4 Flash and V4 Pro, which the company describes as its most powerful open-source models to date. These models show improvements in coding, reasoning, and agentic tasks, which are AI systems capable of planning and executing complex workflows.

The global reach of China's open-source AI strategy is striking. Chinese open-source models accounted for roughly one-third of global AI use in 2025, with DeepSeek emerging as the most widely used among them. This represents a significant shift from the era when Western companies like OpenAI and Google dominated the AI landscape.

"They are basically neck and neck," said Rayan Krishnan, comparing DeepSeek V4 and Moonshot AI's latest model.

Rayan Krishnan, Chief Executive at Vals AI

DeepSeek also expanded its model's context window, allowing it to process longer conversations, documents, and code without losing track of earlier information. The rapid iteration of new AI models makes it difficult for analysts to determine a clear winner in the competition between Chinese and Western systems.

What Would Copyright Reform Actually Look Like?

Policymakers and industry experts are converging on a framework where AI companies pay into an Australian fund that distributes compensation to copyright holders through a government-backed agency. This approach would allow AI training to occur locally while still compensating creators for their work. Public support for this approach is substantial.

• Public Support: A survey commissioned by the nonprofit Good Ancestors and conducted by YouGov found that 61% of respondents supported changing copyright arrangements to enable AI training in Australia, while only 15% wanted laws kept in their current form.
• Economic Scale: McKinsey and Company estimated that encouraging data center investment for frontier AI could generate roughly $80 billion annually in economic activity from 2030 onward.
• Infrastructure Advantages: Australia possesses natural strengths for hosting AI infrastructure, including abundant land, renewable energy resources, political stability, and geographic depth that would position it as a regional compute hub.

The argument is pragmatic. Tech companies can already include most Australian content in their training datasets, which set the neural connections that enable AI models to function. They're simply doing this work in data centers overseas, which means Australia receives no tax revenue, no employment opportunities, and no voice in shaping AI development.

How to Position Australia as an AI Infrastructure Leader

• Establish a Licensing Framework: Create a government-backed system where AI companies pay fees to train models locally, with proceeds distributed to copyright holders and creative industries.
• Leverage Renewable Energy: Use Australia's abundant renewable energy resources to power data centers at lower environmental and operational costs than competitors in other regions.
• Redirect Economic Benefits: Allocate a portion of data center revenue to support creative industries as they adapt to a world where AI-generated content competes directly with human creations.

Why Does Australia's Decision Matter Globally?

Australia's dilemma reflects a broader global tension between protecting creative industries and participating in the AI economy. Content creators, including authors, filmmakers, and musicians, worry that AI training on their work will eventually replace them. That concern is legitimate, but it rests on a false assumption: that restrictive copyright laws will prevent AI companies from using that content.

The reality is that AI advances are continuing regardless of any single country's copyright rules. Chinese companies like DeepSeek are already proving that open-source models can compete with proprietary systems built by Western tech giants. If Australia doesn't participate in building the infrastructure that trains these models, it will simply be left behind while other countries capture the economic benefits.

"Open source is the soft power of technology of the future," said Kevin Xu.

Kevin Xu, Founder at Interconnected Capital

Some of the economic activity generated by data center investment could then be directed to support creative industries, giving them resources to adapt to a world where AI-generated content competes directly with human creations. This creates a pathway where artists and media companies aren't simply protected by law, but actively supported by the wealth generated from AI infrastructure.

The challenge ahead is recognizing that AI-generated content represents a categorically new problem requiring shifts in mindset, laws, and institutions. Whether a novel or song resonates with audiences may matter more than whether it came from authentic human experience, and societies will need to decide what values they want to apply to these questions. But those decisions will be made by countries that are actively building AI infrastructure, not by those trying to keep the technology out.