Darryl van der Peijl
The C-suite wants to know why you’re still running on-premises. Meanwhile, your public cloud costs are spiraling and regulators are asking where your data actually lives. You’re not the only CIO losing sleep over this. In this blog, Darryl van der Peijl, CTO at Splitbrain, explains why the most successful organizations deliberately choose a hybrid approach with most of their data staying local, and how this model minimizes risk while keeping costs predictable.
The end of the cloud-first illusion
Five years ago, everyone said everything should move to the public cloud. Now manufacturing companies are staring at monthly Azure invoices nobody saw coming. Recent research shows that 58% of enterprise workloads run in the public cloud, but the remaining 42% deliberately stays local or hybrid. Why? Because CIOs discovered that public cloud-first without proper planning just means cloud-chaos.
IT leaders in manufacturing and automotive report rising monthly public cloud bills, prompting them to consciously bring workloads back on-premises. Resource-intensive ERP systems and production planning applications are returning first. For manufacturers, latency remains the dealbreaker. A PLC that responds 200 milliseconds late stops the production line. Public cloud simply cannot solve that.
The 90/10 reality
Here’s what successful manufacturing CIOs actually do: they keep the majority of their critical workloads local. Not out of nostalgia, but common sense. That SAP environment running around the clock? Local. The MES controlling your factory floor? Local. Compliance-sensitive HR data? Local.
The answer isn’t more cloud. It’s a hybrid-by-design approach with Azure Local as the foundation. Only truly dynamic workloads belong in the public cloud: development environments, seasonal spikes, AI experiments. Organizations taking this approach report 23% lower operational costs. IBM documented up to 90% lower hosting costs for stable workloads.
CDW, a Fortune 500 IT services provider, made exactly this choice. They migrated over 800 virtual machines to Azure Local within four months. Their non-production workloads now run locally with cloud flexibility. The result: more than 50% lower renewal costs and SQL instances performing nearly twice as fast. Their Director of Platform Engineering sums it up: “Workloads run where they belong. Developers don’t notice any difference between on-premises and Azure. That’s a major win.”
Compliance without the headaches
Regulators want to know one thing: where does your data live, who has access, and can you prove it? With local data, the answer is straightforward. Point to your own datacenter, show your access logs, done.
With fully public cloud-based infrastructure, this gets complicated. Data can be spread across multiple regions, with access rights managed through various identity providers and logging in systems you don’t fully control. Manufacturing CIOs we speak with, report that audits for hybrid setups take days instead of weeks, simply because they can point to their own datacenter.
A hybrid approach lets you consciously choose where data lives. Sensitive business data stays under your direct control. Development and test environments can go to the cloud. This separation makes compliance audits manageable and measurable.
The workload assessment your board will understand
Forget technical exercises. These are the three questions that matter:
Does it run 24/7? Then cloud is probably more expensive than local. Calculate the difference between CAPEX and OPEX over five years.
Does it contain personal data or trade secrets? Then you want control over location and access. Local or sovereign cloud, not public cloud.
Does it need to scale or is it experimental? This is where cloud excels. Development, test workloads, AI labs: perfect for the cloud.
Even with the right balance, two pitfalls remain.
What can go wrong (and how to prevent it)
The biggest risk with hybrid is complexity. Two worlds, double the management, different tools. But platforms like Azure Arc make it possible to manage everything centrally. You maintain on-premises control with cloud flexibility.
At Splitbrain, we specialize in Azure Local implementations for manufacturing companies. We know the pitfalls: from OT/IT integration to identity management. Our approach keeps complexity manageable while ensuring you stay demonstrably compliant.
A second pitfall: vendor lock-in. Once you’re all-in on a single cloud provider, you’re at the mercy of their price increases. With most of your data local, you retain negotiating power.
The financial reality
Terrazone’s five-year TCO analysis shows concrete numbers: for a typical ERP environment (200 vCPU, 200 TB storage), total cloud costs run $854,000 versus $411,000 on-premises. More than double for the same capacity.
For a mid-sized manufacturer, this means:
100% cloud: €2.4 million over five years
100% on-premises: €1.2 million over five years
90/10 hybrid: €980,000 over five years
Puppet’s research confirms that manufacturing and automotive choose this balance for better performance, compliance, and cost control.
Your next step
Public cloud-only strategies are showing their limits. Hybrid by design is the new reality. Azure Local combines cloud innovation with local control. The question isn’t whether you’ll make this transition, but how you’ll find the balance between control and flexibility.
Start with a workload assessment from a business perspective. Identify your 24/7 workloads, your compliance-sensitive data, and your innovation experiments. The optimal distribution becomes clear from there. Oracle research shows that over 60% of enterprise workloads still run on-premises or hybrid. In our practice, we see manufacturing organizations typically landing on a model where 80-90% of stable, critical workloads stay local, with 10-20% in the cloud for dynamic use cases like development and AI experiments.
The real win of hybrid by design isn’t technology. It’s an audit that takes days instead of weeks.
Gerelateerde blogs
