Microsoft, investor BlackRock, and Dubai-based AI investment company MGX are investing 100 billion dollars in AI data centers through the purpose-built Global AI Infrastructure Investment Partnership (GAIIP).
According to the tech giant and both investment companies, the demand for AI capacity is increasing significantly worldwide. This predictably means the demand for data center capacity to run those workloads is also growing. Especially since AI tasks such as machine learning and deep learning often require more computing power than traditional data centers can currently offer.
The partners within the GAIIP believe investments in data center capacity are essential, and are joining forces to make that happen. To this end, they initially want to raise 30 billion dollars from private investment capital from the participants, wealthy private individuals, and large companies. Ultimately, this should lead to the goal of 100 billion dollars in investment capital.
Mostly for US-based data centers
The investments will first benefit the construction of new AI data centers in the US, further driving AI innovation and economic growth there. The remaining investments will take place in so-called “partner countries.”
Exactly which countries these are, was not disclosed, but given the involvement of Emirati investment vehicle MGX, they may be in the United Arab Emirates.
Open architecture and ecosystem
In addition, the partnership plans to work toward an open architecture and broad ecosystem for AI data centers. This architecture/ecosystem combo should provide various partners and companies with full and non-exclusive access to data center capacity.
Nvidia has already confirmed that it will support the investment initiative, especially the AI ecosystem, by contributing its significant expertise in AI data centers and factories. The GAIIP plans to further connect with other key companies in the tech industry to establish AI supply chains for both customers and the tech industry itself.
Also read: Energy hungry data centers in AI era puts nuclear back on the agenda