From Cloud-First to Cloud-Right: The Public Cloud Horizon (1 of 7)

Blog series: From Cloud-First to Cloud-Right | Blog 1 of 7
 

Public cloud is growing. Cloud strategy is getting more selective.

Public cloud spending keeps climbing. The hyperscalers keep posting double-digit quarters. Global infrastructure spend is on pace to clear a trillion dollars, and the trajectory beyond that only steepens. None of that is in dispute.

What has changed is the strategic conversation around it.

For the better part of a decade, the prevailing logic was straightforward: migrate to cloud, migrate quickly, and the value will sort itself out. That logic got a lot of organizations into the cloud. It also left many of them without a clear framework for what to do once they got there — how to govern cost, how to evaluate workload fit, and how to make placement decisions that hold up under financial and operational scrutiny.

The shift I see across our customer base — from mid-market to the largest global enterprises — is away from cloud-first as a default posture and toward deliberate decisions about where each workload runs. I started framing this back in 2016 as matching the workload profile to the right execution venue: understanding a workload's performance demands, data gravity, compliance exposure, cost behavior and lifecycle stage, and then placing it where those characteristics actually fit, whether that's public cloud, private cloud, edge or on-prem. 

The industry has since adopted its own terms: strategic workload placement, cloud-smart thinking, the cloud reset — but the underlying discipline is the same. The next phase of cloud is value-driven, not migration-driven.

The maturity curve no one planned for.

If you've been in enterprise IT long enough, you recognize the pattern. A new capability arrives. Early enthusiasm turns into blanket adoption. Then complexity compounds, the bill comes due and the conversation matures.

Cloud followed that arc exactly. First came the enthusiasm — more cloud was automatically better. Then came the cloud-first strategy, where public cloud became the default starting point for modernization. Now we're in a third phase, and it's harder than the first two: making workload-by-workload decisions that balance speed against cost, agility against control, and innovation against operational reality.

The organizations that moved aggressively into the cloud gained real advantages in speed and agility. But I'd argue the ones that will sustain those advantages are the ones now layering in the discipline of where things should actually run — and being honest when the answer isn't always public cloud.

Hybrid by design, not by accident.

There's a persistent narrative that hybrid cloud is what you end up with when you don't finish your migration. That framing misses the point entirely.

The majority of enterprises I work with operate hybrid estates, but not because they stalled. It's because the workload landscape demands it. Every workload carries a different profile, and every profile points to a different execution venue. Some genuinely belong in public cloud for elasticity and rapid iteration. Some are better served in a private cloud, where cost predictability and operational control matter more. Others need to sit close to the data, the edge or core enterprise systems, where latency and data gravity make proximity non-negotiable.

This is also why repatriation is back in the conversation, and it's worth being direct about that. Recent industry surveys describe what amounts to a cloud reset — public cloud is no longer the automatic default, and private cloud is playing an increasingly significant role in enterprise strategy. Repatriation isn't a retreat. It's what happens when organizations stop treating cloud as an ideology and start treating it as an economic and architectural decision. The triggers are practical: egress costs that erode margins, steady-state workloads where cloud economics invert, compliance mandates that constrain data residency, and the simple desire for more control than a shared platform can provide.

The enterprises getting this right aren't making sweeping moves in either direction. They're making deliberate, workload-by-workload calls backed by data — and they're building the governance to keep those decisions current as conditions change.

FinOps changed the conversation. Now it has to change the operating model.

FinOps brought something the cloud conversation desperately needed: financial accountability. For the first time, organizations began tying cloud spend to business outcomes rather than treating infrastructure as an opaque line item. That was a genuine step forward.

But here's what I see in practice: Adoption is uneven, and maturity varies widely. Plenty of organizations have stood up FinOps dashboards without actually changing how decisions get made. The dashboard shows the waste. The harder question is whether anyone in the room has the authority — and the incentive — to act on it.

The organizations I see extracting the most value are the ones that moved past aggregate cost reporting into unit economics: connecting infrastructure spend to revenue, customer experience and operational outcomes at the workload level. That's where strategic placement actually lives. Not in a monthly spend review, but in the architectural and financial frameworks that shape every placement decision before the workload ever moves.

The leadership question has changed.

The old measure was "How cloud-forward are we?" The better measure, and the one I push with every customer I work with, is "How well are we aligning workload placement to business outcomes?"

That reframing has real consequences. It means cloud strategy can't be owned by infrastructure alone. It needs finance, architecture, security and business stakeholders all at the table, who understand what each workload actually means to the organization. Success isn't migration count — it's the resilience, cost efficiency and governability of the overall portfolio.

The next winners won't be the organizations that moved the most workloads. They'll be the ones who placed them best, and who built the operating model to keep optimizing as the landscape shifts

What's ahead in this series…

This is the first of seven posts, and the big-picture focus is intentional. Over the next six posts, I'm going to get much more specific — starting with why cloud-first broke down as a strategy, then walking through the workload profile framework I use with customers to make placement decisions that actually hold up. From there, we'll get into what "hybrid by design" really looks like, when repatriation makes financial sense (and when it doesn't), how to move FinOps from dashboards to real decisions, and ultimately what the new cloud operating model needs to be. 

If any of that sounds like the conversation you're already having internally, stay with me — this is the series for you.