After two years of cloud sprawl, IT leaders are consolidating platforms. Leaders share what the next 18 months of cloud simplification is likely to look like, and which providers will benefit.
Key Takeaways
The numbers that greet most IT leaders when they complete a genuine cloud inventory are, by nearly every account, worse than expected. According to new research covering 340 enterprises with annual IT spend exceeding $50 million, the average organisation now operates 7.4 distinct cloud platforms. That figure has almost doubled since 2022, when the same cohort averaged 4.1 platforms per enterprise. The growth trajectory was not accidental. It was the predictable outcome of four years of incentive structures that rewarded adoption over architecture.
Three forces drove the expansion. First, pandemic-era procurement acceleration removed traditional governance gates, allowing individual departments to subscribe to cloud services without central IT review. A marketing team adopting a cloud analytics platform here, a finance unit deploying a cloud planning tool there: each decision was individually defensible, but the aggregate created an estate that nobody owned in full. Second, vendor growth programmes actively incentivised departmental buyers with free credits, generous pilot terms, and dedicated account teams who bypassed central procurement entirely. Third, the rapid maturation of specialised SaaS platforms meant that best-of-breed alternatives existed for nearly every workflow category, and line-of-business leaders chose them without considering integration costs or security surface area implications.
The hidden costs that accumulated as a result are now materialising on balance sheets. Licencing is the visible line item, but integration maintenance is typically larger. Each platform boundary requires connectors, middleware, or custom API work to maintain data flow, and that integration estate requires ongoing engineering attention as platforms release new versions. Security and compliance costs compound the problem further: each platform expands the attack surface, introduces new identity management requirements, and demands separate audit evidence for regulatory frameworks. One large financial services firm that participated in the research estimated it was spending 2,400 engineering hours annually just maintaining integrations between platforms that had no strategic dependency on each other. At a fully loaded developer cost of $180,000 per year, that represents $216,000 in engineering time spent on accidental architecture rather than product development.
Perhaps most damaging is the cognitive overhead imposed on engineering and operations teams. When infrastructure spans seven or more platforms, staff must maintain context across multiple vendor interfaces, billing models, security frameworks, and support escalation paths. Onboarding time for new engineers increases substantially; institutional knowledge becomes dangerously concentrated; and the mental bandwidth available for strategic architecture work decreases with every platform added. Cloud architects surveyed for the research consistently identified multi-platform management as the single largest contributor to burnout.
Successful consolidation programmes share a structural characteristic that distinguishes them from the efforts that stall: they begin with workload classification, not platform selection. The instinct in many organisations is to start by choosing the two or three platforms they want to retain and then migrate everything to fit. The evidence from organisations that have completed consolidation cycles suggests that approach reliably fails. Workloads that appear fungible in a spreadsheet carry hidden dependencies, performance requirements, data residency constraints, and compliance needs that only surface when migration is underway. The organisations achieving the best outcomes spend the first 60 to 90 days of a consolidation programme classifying every workload before committing to a target architecture.
The governance structure that enables this is a cloud centre of excellence operating with genuine executive authority. The distinction matters. Many organisations have CoE functions in name only, producing guidance that line-of-business teams are free to ignore. Successful consolidation requires a CoE with the mandate to approve or reject new platform subscriptions, a direct reporting line to the CIO, and a seat at the table during departmental budget cycles. Without that authority, consolidation programmes consistently lose ground to organic sprawl even as they proceed: teams migrate workloads off retiring platforms on one side while new platforms are procured on the other.
Vendor relationships during consolidation require careful management. Both AWS and Azure have developed formal consolidation incentive programmes that provide migration credits, dedicated technical resources, and in some cases contractual commitments on pricing in exchange for increased wallet share. These programmes can be substantial: several CIOs interviewed for this report reported receiving credit packages worth between $800,000 and $2.4M from their primary provider in exchange for committing to a target consolidation architecture. Google Cloud is competing on similar terms specifically in organisations where data analytics and AI workloads are a consolidation priority, offering Vertex AI migration support and BigQuery cost optimisation as anchors for broader platform consolidation.
The phased migration approach that minimises disruption treats platforms in three categories: retire, retain, and review. Retire platforms are those with no strategic workloads, limited integrations, and clear migration paths. They are addressed first, generating early wins and freeing budget for the harder migrations that follow. Retain platforms are the two to four that will form the target architecture. Review platforms are the complex cases where workload dependencies, contractual obligations, or regulatory requirements make migration timelines uncertain. Tackling these in sequence rather than in parallel keeps teams focused and prevents the dilution of effort that characterises unsuccessful consolidations.
"The organisations that are consolidating successfully are the ones that did the workload analysis before they picked the platforms, not after. When you start with the platform and work backwards to which workloads fit, you end up with the same sprawl problem, just with different vendors."
Priya Anand, VP Cloud Strategy, Accenture Technology
The 18-month outlook for cloud consolidation is, by most measures, strongly positive for IT leaders willing to begin now. Budget pressure is not abating. CFOs who accepted multi-year cloud commitments during the growth cycle are now scrutinising renewal terms, and the ROI for consolidation at $2.8M average annual savings is straightforward to justify in any capital allocation process. For IT leaders who have not yet commenced a formal consolidation assessment, the practical starting point is a two-week cloud inventory exercise using one of the commercially available cloud management platforms, producing a comprehensive map of every active subscription, its owner, its workload category, and its estimated annual cost. That inventory is the foundation everything else depends on, and the organisations that have it are already six months ahead of those that do not.
Engineering organisations in the top quartile of technical debt management ship features 40% faster than peers, yet most CTOs rank remediation low on their list.
Security-integrated delivery pipelines have compressed breach-to-detection time significantly. Here is the framework leading engineering organisations are using.
