Why is Megaport Buying Latitude.sh for Data Centre Push?
Megaport has recently announced that it has agreed to acquire Latitude.sh, bringing together the Network-as-a-Service (NaaS) specialist with a global Compute-as-a-Service platform delivering high-performance CPU and GPU infrastructure on demand.
The move is positioned to let enterprises deploy and interconnect critical workloads across more than 1,000 data centres in 26 countries over Megaport’s private, high-speed fabric.
Michael Reid, CEO of Megaport, says: “Megaport has long been trusted by the world’s largest enterprises to move workloads seamlessly between data centres and the cloud.
“By bringing Latitude.sh into the fold, we’re extending that promise beyond the network and into high-performance, optimised compute, complementing cloud providers.
“Together, we will not only serve the massive traditional compute market but will also open the door to the explosive AI infrastructure space and the hyper-growth market of inference.”
What this means for data centres
For operators and tenants, the combined platform is intended to streamline how capacity is provisioned and linked within and between facilities. Megaport’s fabric already connects private infrastructure to major clouds and partners.
By adding Latitude.sh’s dedicated CPUs and GPUs on demand, customers will be able to spin up compute and tie it to networks, clouds and ecosystem partners using the same private backbone. The aim is to shorten time to deploy, reduce operational friction and support high-throughput, low-latency paths that many data centre workloads require.
Latitude.sh’s platform includes developer-friendly APIs, predictable billing and the ability to deploy workloads in as little as five seconds.
The company operates in 20 key markets across 10 countries, with an advanced fleet of more than 7,700 servers and dedicated Nvidia AI clusters. It serves over 1,150 customers spanning enterprise applications, high-performance blockchain nodes, SaaS, gaming, adtech and streaming services.
Bringing compute to the interconnection fabric
The rationale centres on uniting network and compute so enterprises can place resources close to users, partners and clouds while maintaining private connectivity. Megaport’s private, high-speed network is designed to give predictable performance and isolation from the public internet.
Pairing that with Latitude.sh’s bare-metal style infrastructure is intended to let customers stand up GPU or CPU instances and connect them to clouds, partners and other data centres over Megaport’s fabric without building new circuits each time.
Gui Soubihe, CEO of Latitude.sh, says: “This is a tremendous opportunity to extend our compute platform, capable of deploying dedicated CPUs and GPUs on demand, into the world’s largest Network-as-a-Service provider.
“The combination of Latitude.sh’s on-demand optimised compute with Megaport’s global private high-speed network will create a cutting-edge globally automated Infrastructure-as-a-Service platform.”
AI and high-performance workloads
The companies say the integration will address rising demand for AI and performance-sensitive services inside and between data centres.
Latitude.sh’s dedicated Nvidia AI clusters support inference, fine-tuning and training workloads. Coupled with Megaport’s fabric, enterprises will be able to place compute where needed, then link it privately to clouds and data sources across more than 1,000 data centres.
By integrating Latitude.sh into Megaport’s network, enterprises will be able to spin up compute and interconnect it with clouds, partners and facilities worldwide using existing Megaport workflows. This is intended to complement public cloud services while giving customers deployment speed and control over locality.
Strategy and timeline
Michael says: “This acquisition marks a new chapter for Megaport.
“We are building an industry-leading platform where network and compute converge globally. This positions Megaport at the heart of the hybrid cloud and AI-driven future.”
The transaction is expected to close on or before 31 December 2025, subject to customary closing conditions and approvals.
