What happened
Anthropic signed a long-term lease with Singapore-based developer Zerra DC for a data centre in Queensland, Australia. This is Anthropic's first such lease in the country. The facility is part of a $30 billion project located more than 200km west of Brisbane. The deal is specifically for running the Claude AI model, not for training new models. This announcement coincides with a heated political debate in Australia regarding whether copyright laws should be adjusted to accommodate AI investment, with the government reportedly considering an opt-out mechanism for content creators.
Anthropic has entered into its first long-term lease agreement in Australia with Zerra DC, a Singapore-based data centre developer. The facility is part of a larger $30 billion project situated more than 200km west of Brisbane in Queensland. According to SmartCompany, the lease is intended to run Anthropic's existing Claude AI model, distinguishing it from facilities used for training new models.
The announcement arrives at a sensitive time for Australian policy, as the federal government is actively debating how to balance AI investment with copyright protections. A leaked Attorney-General document tabled in the Senate suggests Labor is considering an opt-out system where rights holders can prevent their content from being used for AI training but would forfeit potential income from such usage.
Critics, including Greens senator Sarah Hanson-Young and Australian Society of Authors chair Jennifer Mills, argue that this approach undermines the rights of creators. Mills noted that writers are among the lowest-paid creative workers in Australia, earning an average of $18,000 annually, and that the current legal framework does not yet provide the 'strongest possible protection' promised by Prime Minister Albanese.
AI expert Tony Walsh from the University of NSW suggests that data centre investments give AI companies leverage to lobby for favorable copyright outcomes. He argues that while Australia is a 'cheap place to lobby for change,' there is no national security benefit to allowing overseas companies to train models domestically, and the move risks setting a precedent that weakens digital rights protections.
Source details: smartcompany.com.au ↗
Why it matters
This lease marks a significant physical expansion of Anthropic's infrastructure in Australia, signaling a commitment to the region's data centre ecosystem. However, the timing highlights a critical tension in the AI industry: the conflict between the need for massive computational infrastructure and the legal protections for intellectual property. As AI companies seek to host operations in new jurisdictions, they often lobby for regulatory environments that favor their operational needs, such as relaxed copyright rules for training data. Australia's position as a potential hub for AI infrastructure makes its regulatory stance on copyright a precedent that could influence global AI governance and the rights of creative workers.
The lease represents a concrete step in the global expansion of AI infrastructure, with Anthropic establishing a physical presence in a new major market. This expansion is not just about compute power but also about strategic positioning in a jurisdiction with a growing tech sector.
The intersection of infrastructure deals and copyright policy is a defining feature of the current AI landscape. As AI models require vast amounts of data for training, the legal status of that data becomes a primary point of contention. Australia's potential shift toward an opt-out model could signal a broader trend of regulatory relaxation to attract AI investment.
For creative industries, the outcome of this debate will determine whether their work is treated as a protected asset or a raw material for AI development. The pressure from major tech companies to relax copyright laws poses a significant risk to the economic viability of independent creators and small businesses.
The situation also highlights the geopolitical dimensions of AI development. As nations compete to host AI infrastructure, they may be tempted to offer regulatory incentives that compromise existing legal frameworks. Australia's response will be closely watched by other countries considering similar AI strategies.
What to watch next
Monitor the Australian government's final decision on the proposed copyright opt-out mechanism for AI training. Watch for further infrastructure announcements from other major AI labs in Australia, which may follow Anthropic's lead. Observe the reaction from creative industries and copyright advocacy groups in Australia, as their lobbying efforts could shape the final regulatory framework. Track whether this lease leads to any specific changes in how Anthropic handles data sovereignty or local content usage in its operations.
The final legislative outcome of the Australian copyright debate, particularly regarding the opt-out mechanism for AI training data. Any changes to the law will have immediate implications for how AI companies operate in the country.
Further announcements from other AI companies regarding data centre leases or training facilities in Australia. If Anthropic's move is followed by similar deals from OpenAI or other major players, it could accelerate the regulatory pressure on the government.
The response from international copyright bodies and creative unions to Australia's potential policy shift. A change in Australian law could influence copyright discussions in other jurisdictions, potentially leading to a global race to the bottom in creator protections.
Any operational details released by Anthropic regarding the use of the Queensland data centre, such as whether it will host any training workloads or strictly inference services. This distinction will be crucial for understanding the company's long-term strategy in the region.