AI inference surge strains global data center capacity as cloud and colo builders race to expand

AI inferencing is increasing demand for cloud infrastructure and data center colocation, with companies such as Meta ramping up CapEx spend to rapidly grow the data center footprint.

Crypto companies are moving into the data center business, and Core Scientific hasn't stopped adding capacity even after a failed acquisition bid from CoreWeave.

Web scale and managed infrastructure providers including DigitalOcean, Akamai and Fastly are growing as AI-fueled demand drives momentum with Akamai using its acquisition of Linode expanding compute infrastructure.

Megaport acquired Latitude.sh as well in terms of connectivity and compute infrastructure, in Akamai-like fashion.

AI inference demand is opening up opportunities for distributed-network platforms, like Akamai, Fastly and Megaport, with Akamai having recently unveiled its NVIDIA-powered Inference Cloud.

Rackspace has stabilized its revenue in recent quarters and made some progress, while global expansion continues with Microsoft, Equinix and Global Technical Reality spinning up services in new regions.

The industry is focusing on optimizing operations and builds with AI, addressing challenges like workforce and tool shortages.

Recent updates include Microsoft's AI infrastructure investments in Thailand and the UAE, Equinix's new data centre in Mexico and Global Technical Realty's expansion in London.