Major AI companies OpenAI and Anthropic are expanding their infrastructure strategy to include smaller data centre deployments as the artificial intelligence computing race intensifies. According to reports from CNBC, the two companies have signed major infrastructure agreements over the past year involving facilities with capacities running into hundreds of megawatts and gigawatts. Now, they are exploring 20-30 megawatt capacity deployments to support faster AI deployment requirements. As one source familiar with the discussions told CNBC, both companies have also been involved in talks over US capacity deployments of a similar scale.
Both companies are actively pursuing smaller data centre opportunities across key European markets. As reported by CNBC, Anthropic has held discussions for 20-30 megawatt capacity across the UK and Nordic countries, with four people familiar with the talks confirming these discussions. OpenAI has also explored similar opportunities in the Nordic region, according to two sources cited in the report. An OpenAI spokesperson told CNBC that the company is building a diversified compute portfolio to meet rising AI demand globally, stating "Different workloads need different infrastructure," and adding that OpenAI evaluates opportunities based on requirements including performance, reliability, timing and cost.
The shift toward smaller data centre deployments reflects the evolving nature of AI computing requirements. According to CNBC reports, the demand for faster AI deployment is driving this strategic change, with companies seeking more agile and flexible infrastructure solutions. This represents a significant departure from the traditional approach of securing massive computing capacity through large-scale data centre agreements. Smaller capacity deals can allow AI companies to get access to usable computing infrastructure faster, particularly when larger data centre projects face lengthy development timelines.
The growing focus on smaller facilities is linked to the changing nature of AI workloads, with inference becoming increasingly important as AI computing moves from model training to serving applications in production. According to a JLL report cited by CNBC, inference accounted for 9% of global data centre workloads in 2025, compared with 14% for training. Inference is projected to account for 37% of data centre capacity by 2030, while training is expected to account for 13%. Nvidia also announced in February that it would work with several data centre stakeholders to study smaller facilities designed for distributed AI inference.
Both OpenAI and Anthropic typically rent computing capacity from data centre operators and neocloud providers, while also pursuing large, long-term infrastructure agreements. Anthropic recently signed a roughly $45 billion cloud deal with Nscale, under which it is expected to rent around 460 MW of computing capacity at a data centre development in West Virginia, according to people familiar with the matter cited by CNBC. This demonstrates the companies' dual approach of securing both large-scale and smaller-scale computing capacity to meet varying AI workload requirements across different stages of model development and deployment.
