Multi-cloud is having a moment. Every conference, every vendor briefing, someone’s evangelising the benefits of spreading workloads across AWS, Azure, and GCP. Don’t put all your eggs in one basket. Avoid vendor lock-in. Best-of-breed from each provider.

For hyperscalers with dedicated platform teams, it can work. But for mid-size Australian companies — 200 to 2,000 employees, an IT team of 10 to 30 — multi-cloud is frequently a trap.

How Companies End Up Here

Most mid-size organisations don’t deliberately choose multi-cloud. They drift into it.

The dev team picked AWS because that’s what they knew. Marketing bought a SaaS product on GCP. Finance signed up for Azure because it came bundled with the Microsoft Enterprise Agreement. An acquisition brought in a product on a fourth provider.

Suddenly you’re multi-cloud. Not by design, but by accident. And costs are piling up in places nobody expected.

The Costs Nobody Talks About

Skills multiplication. Each provider has its own networking model, identity system, storage options, and security controls. Different terminology for similar concepts. Your team now needs competence across all of them.

Large enterprises can have dedicated AWS and Azure specialists. Mid-size companies expect the same five people to be expert across multiple platforms. What you actually get is people who are mediocre at all of them instead of strong on one.

Networking complexity. Connecting workloads across providers means transit gateways, VPN tunnels, or third-party overlays. Each is a failure point and a security boundary.

I worked with a Melbourne SaaS company spending $14,000 per month on cross-cloud data transfer. They’d put application servers in AWS and analytics in GCP because “GCP has better data tools.” Technically true. But the transfer costs wiped out any savings.

Security and compliance overhead. Each provider has its own security model, logging format, and compliance certifications. Consistent security posture across platforms means paying for a third-party CSPM tool or manually correlating data from different sources.

For organisations with Australian data residency requirements, multi-cloud adds another layer. You verify controls on each platform independently, and the compliance evidence is platform-specific.

Tooling sprawl. Infrastructure-as-code on AWS means CloudFormation or Terraform with the AWS provider. On Azure, ARM templates or Bicep. GCP has its own deployment manager. Your “portable” Terraform code is really three separate codebases with shared syntax.

Same for monitoring. CloudWatch, Azure Monitor, Google Cloud Monitoring — all doing similar things differently. Most multi-cloud shops buy a third-party observability platform to unify everything. That’s another $50K to $150K per year.

Negotiating weakness. People assume multi-cloud gives you bargaining power. In practice, switching costs are so high that the threat is empty, and providers know it. Worse, splitting spend means you lose volume discounts. A company spending $500K on one provider gets better rates than one spending $200K on AWS and $300K on Azure.

When It Actually Makes Sense

I’m not saying it never works.

Genuine regulatory requirements. Some government contracts mandate specific providers. That’s a real constraint.

Acquired companies. Migrating a stable, self-contained acquired product to your primary provider might not be worth the risk.

Truly provider-specific services. If you need BigQuery for a specific analytics use case and everything else runs on AWS, that’s pragmatism, not ideology.

What I Recommend Instead

For most mid-size organisations: pick one primary cloud provider. Go deep. Build real expertise. Your concentrated spend gives you genuine bargaining power.

Some SaaS products you use will run on different clouds. That’s not multi-cloud strategy — that’s modern business. You don’t manage those platforms.

If you’re already accidentally multi-cloud, map every workload to its provider. Calculate total cost including operational overhead. You’ll probably find consolidating saves more than the switching cost.

Don’t let a vendor convince you that multi-cloud is sophisticated and single-cloud is risky. The riskiest thing a mid-size IT team can do is spread itself too thin across platforms it can’t adequately secure, monitor, or operate.

Depth beats breadth. Every time.