Meta Platforms has significantly expanded its artificial intelligence (AI) infrastructure plans, increasing the capacity of its Hyperion data centre in Louisiana to five gigawatts of compute power as global technology companies intensify investments in the computing infrastructure needed to develop next-generation AI models.
The expansion raises Meta’s total investment in the Richland Parish project to more than $50 billion, underscoring the growing competition among major technology firms to secure the massive computing resources required to train increasingly sophisticated AI systems.
The company said the facility’s planned compute capacity has been increased from an earlier target of two gigawatts to five gigawatts, making it one of the world’s largest AI data centre campuses.
The project forms part of Meta’s broader strategy to strengthen its AI capabilities amid intensifying competition with rivals including OpenAI, Google, Anthropic and xAI.
The announcement comes just days after Meta plans to begin production of its in-house AI training chip, codenamed Iris, in September as part of efforts to reduce dependence on third-party chip suppliers and expand its AI computing capacity to 14 gigawatts by 2027.
The rapid expansion of AI infrastructure reflects soaring global demand for the computing power required to train and deploy large language models and generative AI applications, prompting technology companies to commit tens of billions of dollars to data centres, semiconductors and power infrastructure.
Meta Chief Executive Officer, Mark Zuckerberg, has repeatedly identified AI as the company’s top strategic priority, with investments focused on building the computing backbone needed for future AI products and services. The Louisiana expansion is expected to support the company’s next generation of AI models and digital platforms.
The race to build AI infrastructure has shifted competition beyond software development to include ownership of high-performance computing resources, advanced chips, energy supply and hyperscale data centres capable of supporting increasingly complex AI workloads.
Meanwhile, the development is also expected to increase demand for electricity, networking equipment, cooling technologies and advanced semiconductor components, as hyperscale AI facilities consume substantially more power than traditional cloud data centres.
The Louisiana project has also drawn attention to the growing energy requirements of AI infrastructure, just as local utility providers are planning additional power generation to support the facility. At the same time, regulators and community groups have continued to monitor the potential impact on electricity demand and consumer tariffs.
The global surge in AI infrastructure investments shows that expanding domestic data centre capacity, broadband infrastructure, and reliable power systems is essential to help local businesses in developing nations keep pace with the next phase of AI-driven innovation.