Meta Turns to Google Cloud for AI Data Centre Infrastructure
Meta has signed a six-year cloud computing partnership with Google Cloud worth more than US$10bn, a deal that underscores the growing infrastructure pressures of AI development.
The agreement will see Meta tap into Google Cloudâs servers, storage, networking and related infrastructure to supplement its own build-out of hyperscale data centres.
Cloud computing allows Meta to rent computing power and storage from external providers rather than building all capacity in its own data centres.
Meta expands data centre investment
The deal reflects Metaâs efforts to balance the enormous cost of building dedicated AI data centres with the immediate need for scalable computing power.
In July, Meta CEO Mark Zuckerberg confirmed that the company intends to spend hundreds of billions of dollars to construct multiple large AI data centres in the coming years.
Mark explained that Meta had raised the lower end of its capital expenditure forecast by US$2bn, with spending now projected at between US$66bn and US$72bn.
At the same time, Meta is pursuing external funding and partnerships to share the financial burden of AI infrastructure.
Meta disclosed earlier this month that it is offloading US$2bn in data centre assets to external partners. The agreement with Google therefore provides a way to secure critical infrastructure capacity without bearing the full cost of construction.
This hybrid strategy allows Meta to continue advancing its AI ambitions while managing financial risk.
By combining direct ownership of data centres with cloud partnerships, the company gains access to vast computing resources when needed while retaining long-term infrastructure investments.
How Google Cloud benefits from AI demand
For Google Cloud, the deal is another step in capitalising on the AI-driven demand for computing capacity.
The cloud unit already reported a 32% year-on-year revenue increase in the second quarter of 2025, reflecting its role as a key enabler of AI training and deployment.
Thomas Kurian, CEO of Google Cloud, has positioned the division as a central player in supporting customers building generative AI models.
Training large language models (LLMs) requires immense computing resources, often far beyond what even the biggest technology firms can provide internally.
The partnership with Meta follows a similar cloud computing arrangement Google struck with OpenAI earlier this year, further evidence of cloud providers becoming indispensable to AI firms.
Balancing competition and collaboration
The agreement also highlights the complex relationships in the technology sector.
While Meta and Google are competitors in AI development and digital advertising, they are willing to collaborate where infrastructure needs demand it.
Both companies are investing in their own AI models and platforms, yet the sheer scale of computing required for training means partnerships can outweigh rivalry in infrastructure.
In June, Reuters reports indicated that OpenAI would also begin using Google Cloud services, showing that even direct competitors are increasingly interconnected through shared infrastructure arrangements.
This interdependence reflects a broader trend across the industry: as AI models become more sophisticated and resource-intensive, no single company can easily meet the infrastructure demands alone.
Cloud partnerships are becoming a practical necessity, even among rivals.
Implications for the data centre industry
For the data centre sector, the MetaâGoogle deal further underscores the rising importance of flexible, high-density facilities that can accommodate AI workloads.
Hyperscale operators are facing escalating demand for GPUs, power and advanced cooling, while cloud providers are capturing a larger share of that demand through capacity leasing.
Metaâs approach demonstrates how data centre strategies are evolving.
Ownership of mega-scale facilities remains critical for long-term control, but cloud contracts provide immediate elasticity and cost management. This dual model may define how other firms approach AI infrastructure investment in the years ahead.
As the industry continues to grapple with the financial and technical challenges of AI computing, deals like this highlight the vital role of both hyperscale data centres and cloud partnerships in sustaining growth.
