Anthropic is racing to build data centers in Australia and Japan as the AI company's computing needs outpace its existing infrastructure by 50 percent.
Anthropic is racing to build data centers in Australia and Japan as the AI company's computing needs outpace its existing infrastructure by 50 percent.

Anthropic is racing to build data centers in Australia and Japan as the AI company's computing needs outpace its existing infrastructure by 50 percent.
Anthropic's push into Australia and Japan for data center capacity signals a strategic shift from renting cloud servers to owning physical infrastructure, as the $965 billion AI company scrambles to close a 50 percent gap between computing demand and supply.
"Growth at this pace places an inevitable strain on our infrastructure," Anthropic said in an April blog post, citing "unprecedented consumer growth" that has hurt reliability and performance.
The company is hiring for 13 positions in its compute division, with eight roles based in Australia or Japan focused on data center engineering, operations and deal sourcing. A London-based data center deal sourcing role offered a salary of between £225,000 and £270,000, reflecting the acute labor shortage in the sector.
Anthropic's infrastructure buildout comes as it prepares for an initial public offering expected as soon as autumn 2026. The company's annualized recurring revenue hit $47 billion in May, more than five times the $9 billion reported at the end of 2025, putting pressure on management to secure computing capacity before listing.
Australia offers Anthropic abundant land, significant renewable energy potential and a stable regulatory environment, according to David Wroe, head of AI and Security Program at the Australian Strategic Policy Institute. The country's membership in the Five Eyes intelligence-sharing partnership with the US makes it a trusted destination for sensitive AI compute infrastructure, particularly after two Amazon data centers were targeted during the Middle East conflict.
The main obstacle to a large-scale buildout in Australia is copyright law, which exposes AI companies to lawsuits from rights holders, Wroe said. Some Australian politicians are campaigning against copyright carve-outs for AI training data, creating regulatory uncertainty for companies like Anthropic that rely on large datasets.
A data center energy role in Australia specifically mentions leading "multi-hundred megawatt procurement efforts," according to the job listing, underscoring the scale of Anthropic's ambitions in the region.
Japan's appeal lies in its political stability, reliable power grid, developed subsea cable infrastructure and technically skilled workforce, said Aalok Mehta, director of the Wadhwani AI Center at the Center for Strategic and International Studies. The Japanese government has shown significant interest in domestic AI infrastructure, with Microsoft committing $10 billion to the country in April and GMI Cloud announcing a $12 billion sovereign AI project in March.
Still, securing power remains the defining constraint on data center growth across Asia-Pacific, said Xiaonan Feng, principal analyst of APAC power and renewables at Wood Mackenzie. "Grid availability is emerging as the defining constraint on data centre growth," Feng said, a challenge that applies equally to Anthropic's Japan plans.
Anthropic's expansion mirrors a broader industry trend. The company has already signed a roughly $45 billion agreement with SpaceX and an $18 billion deal with Akamai, while procuring chips and cloud services from Google. But with only half of its computing needs currently met, the Asia-Pacific push represents a critical step toward diversifying away from US-based cloud providers and securing the infrastructure required to support its next-generation models.
For investors, the buildout signals that AI infrastructure spending remains in its early innings. Anthropic's capital expenditure will pressure margins in the near term, but the company's ability to secure power and physical space in stable, energy-rich jurisdictions could determine whether it can sustain the growth trajectory that supports its nearly $1 trillion valuation ahead of its IPO.
This article is for informational purposes only and does not constitute investment advice.