Megaport has launched a storage product, rounding out a platform that now spans compute, networking, and storage under a single operational model. The Brisbane-based company acquired bare-metal provider Latitude.sh late last year, which added compute to its existing network-as-a-service foundation. The storage launch is the next step in that build-out, giving enterprises a single automated platform across all three layers of IT infrastructure. The initial offering focuses on object storage suited for backup, disaster recovery, and data-intensive workloads, with high-performance block and file storage to follow. There are no egress fees, connectivity runs up to 100 Gbps, and pricing is flat per-terabyte with no minimum commitments.
The infrastructure market that Megaport is entering has a well-established cost problem. Enterprises running distributed workloads across cloud, edge, and on-premises environments routinely encounter unpredictable retrieval and egress charges from the major cloud providers.
Those charges are particularly punishing in disaster recovery scenarios, where organizations discover the real cost of their data architecture precisely when they need fast, affordable access. AI training workloads face a related constraint: the time and cost of moving large datasets from storage to compute creates bottlenecks that are as much commercial as technical.
Megaport's argument is that storage co-located with its backbone and with Latitude.sh compute eliminates those bottlenecks by design, keeping data movement private, fast, and free of consumption-based egress pricing.
The internal logic of the platform is an integration play rather than a product extension. Megaport is not simply offering another S3-compatible object store; it is offering storage that is physically and commercially native to its network fabric. That distinction matters because the value of each component increases when the layers share the same operational model.
A customer using Latitude.sh bare-metal GPUs can feed training data from Megaport Storage at wire speed without crossing a public internet boundary or incurring the pricing friction that comes with inter-cloud data movement. This is a build-versus-buy architecture at the infrastructure level: by owning the network, compute, and now storage layers, Megaport can offer pricing and performance guarantees that assembled multi-vendor stacks structurally cannot.
The staffing model behind the platform is relevant to its competitive positioning. Latitude.sh operates across 24 locations in 10 countries with a global headcount of around 80 people, relying on automation rather than on-site personnel. Megaport CEO Michael Reid has described this as the only realistic way to meet AI infrastructure demand at scale given ongoing technical talent constraints.
That model keeps operating costs low enough to sustain competitive pricing, and it is central to what Megaport is trying to demonstrate: that a vertically integrated, heavily automated infrastructure provider can serve the same use cases as hyperscalers at lower cost and with less operational overhead for customers.
The broader industry context is a gradual consolidation away from assembling separate vendors for networking, storage, compute, and security, toward platforms that unify those layers. Hyperscalers have long offered integrated stacks, but at pricing structures and with data residency and sovereignty constraints that not every enterprise can accept.
Megaport's platform, built on private backbone infrastructure and co-located data center facilities, is positioning itself as an alternative for organizations that want integrated performance without ceding control to a major cloud provider.
