Anthropic has struck a new datacenter partnership with Australian infrastructure heavyweight Macquarie and Singapore’s sovereign wealth fund GIC, according to The Information. The deal pulls two of the world’s largest institutional investors directly into the buildout behind Claude, Anthropic’s family of AI models. It’s another sign that the race for AI compute is now as much a financing story as a technology one.
What stands out here is who’s writing the checks. Macquarie is one of the planet’s biggest infrastructure investors, the kind of firm that funds toll roads, power grids, and data infrastructure. GIC manages a huge chunk of Singapore’s national reserves. When money like that flows into AI datacenters, it tells you these facilities are being treated like long-term infrastructure assets, not speculative tech bets.
Why this matters
Compute is the bottleneck in AI right now. Training and running frontier models takes enormous clusters of chips, power, and cooling, and the bill runs into the billions. No single lab wants to carry all of that on its own balance sheet.
That’s the problem this partnership addresses. By bringing in infrastructure investors and a sovereign fund, Anthropic can secure the capacity it needs without shouldering every dollar of the construction cost itself.
Three things this signals:
- Capital is the new moat. Access to compute increasingly decides who stays at the frontier. Locking in financing partners is a competitive move, not just an accounting one.
- AI datacenters are becoming an asset class. Pension-style and sovereign capital chasing these projects means investors expect steady, long-dated returns from AI infrastructure.
- Anthropic is scaling to compete. The company has been closing the gap with OpenAI, and you can’t do that without a matching hardware footprint.
The context
This fits a pattern that’s been building across the industry. OpenAI has leaned on Microsoft and a web of partners, including the Stargate infrastructure effort, to fund its compute. Anthropic already has deep ties to Amazon and Google, both of which have invested billions and supply cloud capacity.
Adding Macquarie and GIC widens that base. Instead of relying only on Big Tech backers, Anthropic is tapping the same pools of capital that build physical infrastructure worldwide. The Information’s reporting frames this as part of Anthropic’s broader effort to lock down the resources it needs for the next wave of model development.
The status quo before deals like this was simpler: a lab paired with a cloud giant, and the cloud giant handled the hardware. What’s changing is the arrival of dedicated infrastructure and sovereign money as direct partners in the buildout. That spreads the risk and the cost across more balance sheets.
What to expect
For practitioners and anyone tracking the AI market, a few implications are worth holding onto.
- More capacity coming online. Fresh financing means more datacenter construction, which should ease some of the compute crunch over time.
- More of these deals ahead. If infrastructure and sovereign investors are comfortable funding AI datacenters, expect competitors to court the same capital.
- Higher stakes for Anthropic. Big outside money raises expectations. Investors like these want to see the buildout translate into a durable business.
This is significant because it shows AI’s center of gravity shifting toward physical infrastructure and the deep-pocketed institutions that fund it. The models get the headlines, but the datacenters, and who pays for them, increasingly decide who wins.
Anthropic hasn’t laid out the full scope publicly, and the finer terms will matter. For now, the takeaway is clear: the company is stacking up the capital and capacity to keep pace at the frontier. More details are available in the original report from The Information.