Feature
The launch of AWS EC2 in 2006 is widely regarded as the turning point that made cloud computing a practical reality for businesses and developers.
In the near-20 years since, the technology has evolved and improved to support businesses in all kinds of tasks.
In fact, it has gotten to the point where workplace computing and cloud computing have become synonymous — or nearly have.
Whispers of cloud computing "holdouts" — organizations that have stuck with on-premises computing rather than moving everything in the cloud — arise from time to time, usually within the context of how that choice puts them at a disadvantage. Yet new research confirms what many workplace managers have long known, namely, that many companies are still quite happy with on-premises technology.
Broader Cloud Computing Trends
Global technology research and advisory firm ISG’s recent Cloud Market Lens Study unveiled pivotal trends in cloud services spending that are reshaping the IT landscape. As organizations continue to prioritize cloud adoption, these insights can provide strategic direction for IT teams and business leaders. Some of the key findings include:
- 17% of IT budgets are now allocated to cloud services, reflecting the critical role of cloud in modern IT infrastructure.
- 30% of average cloud spend is devoted to software-as-a-service, with infrastructure-as-a-service increasing to 24%.
- Only 19% of budgets are devoted to the private cloud, which marks a decline.
The anticipated rise in AI adoption is likely to boost cloud spending across various sectors, sparking renewed interest in private cloud solutions, the report reads. As businesses increasingly integrate AI and other advanced technologies, the importance of robust cloud infrastructure is expected to grow even more.
The report also noted the growing popularity of multi-cloud strategies. Cloud migration continues at a rapid pace, with the highest adoption seen in ecommerce, ITSM and BI/analytics applications. According to the report the next two years will see enterprises focusing on BI/analytics and AI development.
The Cloud Computing Holdouts
The research also points out that, for the moment at least, 50% of apps are still found on-premises. With cloud development as rapid as it has been over the past few years this is no surprise, said Alex Bakker, distinguished analyst and director of primary research at ISG.
ISG studies have shown that enterprises like the ones studied in this research have an average of 1,600-1,700 applications in their tech stacks, he said. Many of these are legacy, low-volume or non-strategic. Given limited resources, businesses must prioritize and make a specific value judgment about whether modernization — in this case, migrating to the cloud — will help the business achieve something new.
The 50% figure indicates that in many cases it is entirely fine to remain on premises, he said. In part because many workloads have no specific scaling requirements — they have to run, but don’t have variable demand. Old servers are now viable for longer, meaning that on-premises servers can continue to operate for many years even once they are fully depreciated.
Unless an organization sees a business need to move to the cloud, there's little in favor of migrating. "I do not think there is a 'forcing function' for most apps. Some, like SAP, are working to force this change for certain applications," Bakker said.
However, he sees signs that this will change soon. "The focus on cost optimization for cloud expenditure right now will weigh 'default' growth to the cloud,” he added. "But that will be balanced by the new business applications that are built in the cloud — possibly led by the platform services for features like Lambda functions, databases or Generative AI features."
Related Article: Is Your Cloud Service Provider Trustworthy?
Balancing Scalability With Compliance and Client Trust
While it isn't the only reason organizations keep apps on premises, organizations that hold large amounts of sensitive data add another layer of security by remaining on-prem.
Liudas Kanapienis, CEO and co-founder of finance compliance specialist Ondato, said that while they have adopted cloud solutions for certain aspects of their operations, a significant portion of their core technology stack remains on-premises. The primary drivers of this decision were the sensitive nature of the data the company handles and the stringent regulatory requirements they must meet.
“In the financial services and compliance sector, data privacy and security are paramount. Many of our clients, particularly large financial institutions, have strict data residency requirements,” he said.
The result is that they need to guarantee customers that personal and financial information remains within specific geographical boundaries. On-premises infrastructure gives greater control over data location and access.
The introduction of regulations like GDPR and various financial services regulations have further reinforced the need, by necessitating a level of control and auditability that's easier to maintain with on-premises solutions. “We can more easily demonstrate compliance when we have direct physical control over our servers and data centers,” Kanapienis added.
“That said, we're not entirely avoiding the cloud. We use a hybrid approach, leveraging cloud services for less sensitive operations and scaling needs. For instance, our customer support systems and some analytics tools are cloud-based,” he said.
"The decision between cloud and on-premises isn't just about technology — it's about balancing innovation with security, compliance and client trust."
4 Reasons Companies Remain On-Premises
The decision to keep significant portions of technology stacks on-premises, despite the availability and advantages of cloud solutions, can be attributed to several strategic considerations deeply rooted in business operations and objectives, Philip Alves, CEO of DevSquad, added. He points to four different reasons for the decision to remain on-premises:
1. Compliance
Security and regulatory compliance play a crucial role, especially for industries such as healthcare and finance where data sensitivity is paramount, Alves said. On-premises infrastructures offer companies direct control over their security measures, potentially reducing vulnerabilities associated with third-party cloud platforms.
2. Performance Concerns
Performance and reliability concerns also dictate the need for on-premises setups. Certain applications require the lowest possible latency, and proximity to physical hardware can be critical, particularly in sectors like manufacturing and telecommunications, he continued.
Conference
Oct
20
Gartner IT Symposium/Xpo Orlando 2025
Register
Conference
Oct
27
Gartner HR Symposium/Xpo Orlando 2025
Register
“Moreover, the initial investment in on-premises infrastructure for some companies is substantial, making a full transition to cloud not only a technical challenge but also a financial one,” he added.
3. Data Sovereignty
Data sovereignty and control is also becoming increasingly important for companies operating in multiple geographies, he said. Many companies often choose to keep sensitive data in-house to maintain control over data handling and storage, ensuring compliance with national laws and regulations that might restrict data storage to specific geographic locations.
4. Legacy Systems
Integration with legacy systems poses a significant challenge, Alves added. Many enterprises operate on deeply entrenched systems that are difficult or economically unfeasible to migrate to the cloud. The risk of disrupting existing processes and the high costs associated with redesigning these systems for cloud compatibility are substantial deterrents.
Related Article: Why Enterprises Are Bringing Their Workloads to Multi-Cloud Environments
Other Factors in Favor of On-Prem: Network Stability and Egress Costs
For Cassie Fields, president of SaaS company Autoleap, organizations need to consider another factor when debating a migration to the cloud, notably the ability to control latency at a granular level for sensitive workloads.
Here she cites the example of financial institutions that conduct high-frequency trades. In that case, they require on-premises servers because of the consistent delivery in terms of microseconds, which is what they need for precision.
However, Fields said it also comes down to network stability, which is critical in applications where even slight disruptions can cause huge losses.
She also believes there is a problem with legacy systems and what she describes as the sunk cost fallacy, when organizations continue using old systems to justify the past investments.
Such legacy systems may have been designed specifically for certain tasks and therefore cannot be easily migrated into cloud environments, she said. In cases like this it is more than a technology problem, it's a culture one. Staff members who have been supporting these platforms for many years might resist change due to a lack of skills or fear that they will become irrelevant.
Finally, she also points to egress costs in cloud computing, which refer to charges incurred when moving data out of storage facilities provided by service providers such as Amazon Web Services (AWS).
These hidden fees can quickly add up, especially if one deals with large volumes of information, thus making it expensive for organizations running heavy data applications to shift them from on-premises locations into public clouds.
"This approach allows enterprises to keep track their operational expenses better since they know what exactly goes where at all times, while saving money too. It balances innovation against fiscal prudence so that businesses do not get caught up by fine print associated with using cloud services,” Fields concluded.