The path to SAP S/4HANA has long felt like a one-way street leading directly to the hyperscale public cloud for most enterprise leaders. However, according to Jim Loiacono, Sales Lead, HPE Worldwide SAP Private Cloud, that assumption might be the single greatest barrier to actual business value. Loiacono recounts a recent experience with a Texas-based customer who spent nearly two years navigating Cloud ERP conversations before finally hearing an offer that resonated with its actual operational needs. The company’s ultimate choice? Cloud ERP Private but hosted by SAP within the customer’s own data center.
“If I could offer SAP customers only one statement of advice, it would be that SAP customers should evaluate all deployment options for Cloud ERP,” Loiacono says.
He adds that the industry often conflates Cloud ERP with Data Center Migration, but the two are not inextricably linked. “While moving to Cloud ERP requires a shift from perpetual licensing to a subscription model and sharing database administration responsibilities with SAP, it does not mandate moving your systems away from where they currently run—whether that is on-site or in a colocation facility,” Loiacono explains.
The Iceberg of Integration
One of the primary reasons for resistance to Cloud ERP adoption is the incorrect assumption that migration to a hyperscaler is mandatory. This misconception creates unnecessary risk and business disruption. SAPinsider recently reported that the top two reasons for hesitating on SAP Cloud ERP Private are the perceived value and the total cost. Loiacono argues that mitigating both cost and change actually results in higher value.
However, to understand the real cost, one must look beneath the surface of the Cloud ERP label. Enterprise-scale SAP customers understand that ERP is just one critical organ in a much larger digital body. Loiacono uses a sharp analogy: “These application dependencies can create differences in perceived value like a waiter’s perspective versus a head chef’s perspective of what is happening in a restaurant kitchen—the waiter has a partial perspective, and the head chef has a full picture.”
Without proper due diligence, decision-makers—the waiters in this scenario—may fail to realize that high-level subscription and migration estimates represent only the tip of the iceberg. They often significantly understate the total cost of a holistic migration by ignoring customer-specific dependencies like custom code, multi-level integrations, and linked reporting with non-SAP systems. Loiacono points to a CFO whose holistic ERP ecosystem was “limping along” after a supposedly successful SAP implementation because the migration team was unaware of these critical ecosystem dependencies.
The Physics of Data and the AI Factory
According to Loiacono, we have transitioned from a Cloud First era into a Data First world. While a single-hyperscaler strategy might work for smaller firms, modern enterprise-scale customers consume a mix of public, private, and on-site applications. Limiting an ecosystem to one cloud provider is not only uncommon but can lead to high-priced vendor lock-in.
“The real challenge today is data gravity,” Loiacono says. “Aggregating and analyzing insight from disparate datasets is compounded by hyperscale cloud movement fees—egress, inter-availability zone, and inter-region costs—and the simple laws of physics.”
For example, downloading one petabyte of data at a sustained throughput of 10Gbps would take approximately nine days and cost hundreds of thousands of dollars in egress fees every single time. Conversely, uploading that same data is slow and often non-compliant with organizational privacy rules.
This is where the SAP Cloud ERP Private, Customer Data Center (CDC) option becomes a strategic lever. It allows organizations to eliminate data movement costs by keeping the ERP processing power where the data generated lives. Loiacono believes that while SAP’s Joule AI will be a vital agent for ERP-specific tasks, most enterprises will maintain a locally managed AI factory for aggregated inference. This prevents limiting employee productivity based on the credit-based pricing models common in public clouds. As Satya Nadella, Microsoft’s CEO stated at the 2026 World Economic Forum, a company’s technical sovereignty lies in its firm control of its data and its ability to mix and orchestrate different artificial intelligences.
The Sovereignty Illusion
Data privacy and control have moved from legal footnotes to board-level priorities Geopolitical tensions and recent service-denial sanctions have highlighted the risks of third-party hosting. In 2025, an Indian petroleum corporation, Nayara Energy, was suddenly denied access to its services due to EU sanctions without prior notice. Similarly, the chief prosecutor of the International Criminal Court lost access to email due to US sanctions.
Loiacono notes that many customers were hesitant to adopt Cloud ERP until they were offered a sovereign option in their own data centers. While US hyperscalers have created Sovereign Cloud subsidiaries, legal experts like Dave Michels and French MP Philippe Latombe argue that these measures may offer only a sovereignty illusion. Because of the U.S. FISA Section 702, cloud providers are mandated to cooperate with security agencies globally and are prohibited from even notifying the customer of the data acquisition.
“The ultimate data sovereignty of SAP ERP data results from a customer maintaining their own direct physical control of such data in their own data center,” Loiacono asserts. Options like SAP Cloud ERP Private CDC and SAP Sovereign Cloud On-site allow systems to continue functioning even if temporarily disconnected from the outside world, providing the ultimate level of protection against external service denials.
The Economic Reality Check
When evaluating the business case, the math often shifts based on the size of the database and the duration of the contract. For smaller ERP databases, hyperscalers generally charge less due to shared components. However, for longer-duration contracts and larger databases, hyperscale providers generally charge more for Infrastructure as a Service (IaaS) than on-site IaaS cloud services.
Moreover, public cloud providers charge for every data movement between availability zones and regions. In contrast, on-site data centers and colocation providers typically allow for unlimited data transmission within the purchased network bandwidth. This also applies to AI: hyperscale AI generally charges per inference token or credit, whereas self-hosted AI can provide for unlimited inference once the infrastructure is in place.
Actionability: Aligning with Strategy
Loiacono asserts that the goal for SAP professionals should be to align Cloud ERP with their existing business strategy, not the other way around. “SAP intentionally created these flexible deployment models so that customers can remain in stride with SAP’s clean core strategy without resetting their entire data management strategy to zero,” he notes.
His advice is prescriptive: Explicitly ask SAP to evaluate on-site deployment options. These options ensure that no SAP innovation is sacrificed—you receive the same Private edition features and Joule AI capabilities—while maintaining years of purpose-built efficiency and security within your own walls.
“SAP intentionally created flexible deployment options,” Loiacono concludes. Whether you are optimizing for AI inference, protecting sensitive organizational secrets, or simply avoiding the iceberg of hidden migration costs, the most efficient path forward is the one that recognizes your data as your most valuable asset.
HPE
Sponsored by:
