Ask a customer to compare two cloud providers and the first thing they will build is a spreadsheet. Virtual CPUs, memory, storage, price per hour. The cheapest row usually wins.
That is the trouble with selling infrastructure alone. It is the part of cloud that is easiest to compare, and so the hardest to earn a margin on. Global providers, regional clouds, data centres and MSPs all sell compute. When the product is identical on paper, the conversation ends at price.
The providers that build durable businesses also sell things a customer cannot put in a spreadsheet, because what the customer is buying is an outcome.
Infrastructure is the floor, not the ceiling
None of this means infrastructure does not matter. Compute, storage and networking are the foundation. They bring customers in and they produce steady monthly revenue. But competing only at that layer means competing on capacity, availability and price against companies whose scale you will never match.
A more useful way to see a cloud business is in layers. Each one up is harder to compare, closer to the customer’s real problem, and further from a price war.
Infrastructure is what customers compare. Everything above it is what they choose.
Platform services: selling time
A customer running PostgreSQL on a virtual machine is buying capacity. A customer buying a managed database is buying back the hours their team would have spent patching, backing up and tuning it. The same is true of Kubernetes, object storage and container platforms. These services are priced against the effort they save rather than the hardware they use, and they tie a workload more closely to the platform it runs on.
Continuity: selling against a risk
Backup and disaster recovery are bought for a different reason: because something might go wrong. So they are priced against the cost of an outage, not the cost of storage. They also sell naturally in tiers. A development environment and a core banking system do not need the same recovery plan, and a provider that offers several tiers of protection can match the price to the stakes.
Security and compliance: selling evidence
Regulated customers in banking, healthcare, government and telecoms need more than controls. They need to prove the controls exist, to auditors, to regulators and to their own boards. Identity and access management, key management, privileged session recording, audit trails, compliance dashboards and regulatory reporting can all be sold as services.
One word of care. A provider cannot make a customer compliant and should never sell as though it can; certification attaches to the organisation being assessed. What a provider can sell is the part that is expensive to do alone: controls the platform enforces, and evidence produced automatically rather than assembled by hand before every audit. That turns compliance from a cost the customer carries into a service the customer buys.
Managed services: selling expertise
Many organisations can buy cloud. Fewer can run it well. Monitoring, administration, migration, optimisation and application support earn revenue from expertise rather than hardware. This is also the layer where a local provider’s advantages are hardest for a distant one to match: engineers in the same time zone who already know the customer’s environment.
AI: selling what is scarce
GPU capacity, AI workspaces, model hosting and development environments are in demand precisely because they are scarce and hard to build. For regulated customers they come with a condition: training data and models often have to stay under local governance. A provider that can offer AI infrastructure inside the customer’s own jurisdiction is selling something few competitors can.
Sovereignty and specialisation: selling fit
The top layer is fit. A cloud designed for banks, hospitals or government is not infrastructure with a different logo. It arrives with the controls, deployment models and support those sectors expect. Sovereignty belongs here too. Local hosting, local operations and local governance are not a compliance tax on a provider. For the customers who need them, they are the reason to buy.
Depth, not volume
Put the layers together and the measure of a cloud business changes. The question is no longer only how many customers you have, but how many things each customer buys from you. A customer with virtual machines can leave when a cheaper offer appears. A customer with virtual machines, managed databases, a recovery tier, security services and an AI workspace has a relationship with you, and every layer makes the next one easier to sell.
Some of those layers you will build and run yourself. Others you will sell from partners through a marketplace. Either way, the difficulty has always been the same: each layer is its own engineering programme, and most providers never get past the first two.
One platform for every layer
Oasis Stack puts the layers on one platform, above infrastructure the provider already owns. Cloud Foundation covers compute, Kubernetes, managed PostgreSQL, MySQL, MariaDB, MongoDB and Microsoft SQL Server, and backup and disaster recovery across tiers 0 to 3. Governance, Security & Compliance brings identity, policy and audit, hardware security modules, bring-your-own keys, a two-person rule and session recording. Infrastructure for AI provides GPU services, AI Factory and WorkspaceAI. Digital Transformation adds application modernisation and virtual desktops. CloudSouq sells the partner services you do not run. Oasis Core carries every one of them through the same lifecycle, from order to invoice.
What you charge for each layer is yours to decide. The platform is a flat fee per deployment, banded by scale, so the margin you build above infrastructure stays with you.