1300 166 138 Tell us what you need
News / Anthropic

Anthropic opens Sydney office under National AI Plan MOU

By Monika Tantau • Published 7 months ago
Anthropic's new Sydney office and the first MOU under Australia's National AI Plan

Key takeaways

• Anthropic has opened its first Australian office in Sydney to serve the A/NZ market
• Theo Hourmouzis, formerly of Snowflake, has been appointed as the first ANZ general manager
• The company signed the first MOU under Australia's National AI Plan with the AI Safety Institute
• Australia ranks in Anthropic's top five countries for per-capita Claude usage, concentrated in NSW and Victoria

Most Read

Anthropic has formally opened its first Australian office in Sydney and named Theo Hourmouzis, previously Snowflake’s senior vice president for Australia, New Zealand and ASEAN, as its inaugural general manager for Australia and New Zealand, according to ARN. The move follows a memorandum of understanding signed in April with the federal government — the first such agreement under Australia’s National AI Plan.

The opening reflects a broader push by US AI vendors to put senior leadership and physical infrastructure in front of Australian enterprise and government buyers. With OpenAI, Microsoft and Google all scaling regional presence, Anthropic’s Sydney play sets up direct competition for ANZ AI spending across financial services, public sector and professional services.

What Anthropic announced

Hourmouzis joined Anthropic from Snowflake, where he led operations across Australia, New Zealand and Southeast Asia. His earlier roles include leadership positions at Sysdig, Cohesity, Nimble Storage, NetApp, Thomas Duryea Consulting, Hansen Technologies and Pearson Australia Group — more than two decades of regional experience.

He was scheduled to meet partners and customers in Sydney alongside visiting Anthropic executives the week of the office opening. The Sydney base will house a local team focused on enterprise sales, partner enablement and government engagement. Anthropic framed the appointment as a recognition that Australian organisations want regional support for AI deployment, particularly in regulated sectors. The vendor has not disclosed headcount targets.

The MOU and Australia’s regulatory posture

The agreement signed with the federal government in Canberra was the first MOU executed under the National AI Plan. Under its terms, Anthropic will share frontier model risk data with Australia’s AI Safety Institute and provide adoption metrics from its Economic Index to inform government policymaking.

The deal sits within Australia’s broader stance of preferring voluntary bilateral arrangements with major AI vendors over comprehensive mandatory guardrails. That approach has divided industry and labour groups domestically, with the Albanese government continuing to weigh whether existing legislation is adequate or whether overarching AI regulation is required, as reported by The Age.

For Australian organisations, the MOU has practical implications. Vendors with formal government engagement now operate in a different procurement category from those without. Risk teams comparing AI suppliers will weigh the transparency commitments embedded in such agreements, particularly in financial services, defence and health, where APRA, the OAIC and sectoral regulators are watching how AI capabilities are evaluated and deployed.

Existing Australian customers and research ties

Anthropic disclosed an existing Australian customer base that includes Commonwealth Bank and Quantium, alongside platform partnerships with Canva and Xero. Local research collaborations span the Australian National University, the Garvan Institute of Medical Research, Murdoch Children’s Research Institute, Curtin University and Swinburne University. Not-for-profit partnerships were announced with YMCA South Australia.

Australia ranks in the top five countries globally for per-capita use of Claude, Anthropic’s chatbot and developer-facing model. Usage is concentrated in NSW and Victoria, mirroring the geographic distribution of professional services and financial firms that have driven enterprise AI adoption since 2024.

For organisations evaluating conversational AI deployments, the increased local presence means access to direct vendor support and tighter feedback loops on model behaviour. Where Australian businesses previously had to escalate through US-based account teams, they should now expect regional sales and solutioning resources within Anthropic itself.

Australian business leaders weighing AI vendor competition and governance posture under APRA CPS-230

Why Australian businesses should pay attention

Three implications matter for Australian operators.

First, vendor competition intensifies. With Anthropic, OpenAI, Microsoft and Google all running serious regional plays, procurement leverage is improving. Mid-market and enterprise buyers should benchmark commercial terms — pricing, data residency, and SLA commitments — across at least two vendors before signing multi-year deployments. A breakdown of Claude versus ChatGPT for business workloads is useful for narrowing technical fit before negotiating commercials.

Second, governance posture becomes a differentiator. The MOU links Anthropic to the AI Safety Institute and ties its narrative to responsible deployment. Boards and risk committees should expect closer scrutiny of vendor governance — both their own and their suppliers’ — in 2026 and 2027, particularly under APRA CPS-230 for financial firms and the forthcoming Privacy Act amendments.

Third, with Claude usage already concentrated in Australia, organisations that depend on AI search and answer engines should track how Anthropic’s models cite content. Visibility in AI search results — including Claude.ai answers — is becoming a meaningful traffic source for B2B publishers and service businesses.

A new front in the enterprise AI race

Anthropic’s Sydney move closes the gap with rivals that established Australian operations earlier. Microsoft has run an Azure region in NSW for years; Google Cloud operates Sydney and Melbourne regions; OpenAI signed its own Australian government engagements in 2025. The competitive question is no longer whether US AI vendors will commit to Australia — it is which can credibly serve enterprise governance, integration and support requirements at scale.

For Australian businesses, the practical consequence is choice. With four credible enterprise AI vendors operating locally — and emerging local alternatives to consider — the cost of diligence is going up but so is the option value. Organisations that lock in single-vendor commitments before assessing this newly competitive landscape may find themselves rebuilding integrations sooner than they would prefer.

Scroll to Top