The Right Cloud, on Our Terms

Vriti Magee | Jul 9th 2025

“ERP designed for resilience, governed on-prem, and architected for the realities of critical infrastructure” Illustrated by ChatGPT & DALL·E

Why one of North America’s largest pipeline operators chose SAP Cloud ERP on-prem—without compromising sovereignty, resilience, or enterprise scale.

Reflections from Cloud Field Day23, on how a critical infrastructure enterprise redefined cloud strategy on its own terms.

Few organisations operate at the scale—or complexity—of Energy Transfer —one of the largest midstream pipeline operators in North America, responsible for moving close to a third of the continent’s hydrocarbons. With more than 125,000 miles of pipeline infrastructure, and subsidiaries like Sunoco LP—projected to reach $50B in revenue—contributing to an enterprise value approaching $150 billion, the company operates at an intensity that stretches what most IT strategies are built to handle.

So when Energy Transfer made the decision to modernise its SAP estate, it wasn’t about following a trend. It was about controlling risk, maintaining precision, and preparing for innovation—on their own terms.

At Cloud Field Day23, during HPE’s joint session with SAP, Randall Grogan, Senior Director of IT SAP & Financial Applications, explained why Energy Transfer chose to adopt SAP Cloud ERP through the Customer Data Center (CDC) model—instead of migrating to a hyperscaler, as many are encouraged to do.

This wasn’t a rejection of cloud. It was a recalibration—toward a version of cloud that could meet them where they were.

🔄 A Rational Reset

Five years ago, Energy Transfer undertook a substantial application rationalisation. Over 50 fragmented systems—including multiple SAP ECC instances—were consolidated into a single instance of SAP S/4HANA. It was, by energy industry standards, a bold move made early. And yet, it was also just the beginning.

As SAP began encouraging organisations to move toward its Cloud ERP offering—formerly known as RISE—the question became not whether to move, but how.

For many enterprises, hyperscaler clouds are an obvious choice. But for Energy Transfer, with its deeply integrated oil and gas systems, real-time data streams, and critical infrastructure status, the cloud had to be approached differently.

🧱 Building Around Constraints, Not Against Them

The CDC model they chose isn’t just a technical option—it’s a strategic posture. HPE provides the infrastructure, SAP manages the environment, but the systems run inside Energy Transfer’s own data centers, in Houston and Dallas.

Why?

Latency: ~98% of SAP transactions flow from oil and gas-specific platforms. Externalising compute would have introduced friction where there can be none.
Cybersecurity: They are a critical infrastructure operator. “We feel like we have more security within our own data center than we would in the hyperscaler,” Grogan shared—citing the Colonial Pipeline incident as a galvanising example.
Cost: Their data centers already serve third parties and internal groups. There was no opportunity cost to staying local—and no financial case for moving to hyperscale.
Governance: Their infrastructure and cybersecurity teams are in place. Why give away operational sovereignty?

📶 Cloud Isn’t One Thing

What struck me most was the nuance in Grogan’s perspective. This wasn’t cloud evangelism—or cloud scepticism. It was cloud realism.

As the HPE team described it, Energy Transfer’s approach reflects “the right cloud for the right workload at the right time.”

The CDC model allowed them to unlock SAP’s innovation roadmap—including future access to AI capabilities—without sacrificing architectural integrity.

And it keeps options open. SAP described how some customers are experimenting with smaller “user box” deployments to test agentic AI features and involve the business side in what used to be a purely IT domain.

🤖 AI Without Abandonment

There was no sweeping promise that AI would revolutionise Energy Transfer’s operations overnight. Instead, Grogan expressed something more measured—and more believable:

“We don’t know exactly what AI is going to bring for us—but we believe it does have value.”

The SAP AI engine will operate on ERP data. Whether proprietary operational datasets are also used is up to each company. Governance stays local. Flexibility stays intact.

And when AI maturity converges with operational readiness, they’ll be prepared—because the groundwork was already laid.

🔐 Shared Infrastructure, Shared Responsibility

In our discussion, it was clear this model doesn’t relieve internal teams of responsibility—it reframes it. There are still challenges ahead. Coordinating SAP, HPE, internal teams, and external migration partners is no small feat. Governance continuity—from pre-sales through to go-live—is treated as non-negotiable. Data backups, DR procedures, ransomware readiness—they all remain Energy Transfer’s concern.

If anything, the company’s layered approach—shared operational responsibility, strong internal infrastructure, and a culture of engineering precision—makes the complexity manageable.

🪶 The Weight of the Invisible

If there was a single takeaway from that session, it’s this: modernisation doesn’t always look like a leap.

In Energy Transfer’s case, it’s a deliberate recalibration—a move that preserves performance while opening the door to innovation. A model that says “yes” to AI without saying “no” to control. From maintaining infrastructure to preparing for innovation. From application sprawl to architectural clarity. From reactive defence to deliberate design.

From the outside, it might look like Energy Transfer hasn’t moved. In reality, they’ve shifted the centre of gravity—without compromising the perimeter.

This is why CDC is not a fallback—it’s a forward move with guardrails.

And in today’s enterprise, that may be the boldest move of all.

📄 Links for Further Reading and Reference

To explore more on the ideas discussed in this article: