Is on-premise warehouse software making a comeback?
Over the past decade or so, we’ve seen a shift away from on-premise warehouse software toward cloud-based alternatives, as operators look for more scalable solutions without the need to invest heavily in perpetual licenses and server infrastructure. One of the big barriers to adoption in the early years of cloud WMS solutions was concern over cybersecurity and data governance. However, in recent years, these concerns have been largely addressed, at least outside of some highly-regulated industries that still favor on-premise deployments.
One of the main types of warehouse software where on-premise deployments still dominates is warehouse control software (WCS). This requires ultra-low latency to control the fine movements of high-throughput automation in terms of milliseconds. However, we’ve heard of several companies that have deployed WCS solutions in the cloud without any issues around latency. Between 2018 and 2025, the share of SaaS deployments across all warehouse software types increased from just over 30% to more 55%.
I mention this because it seemed as though the writing was on the wall. The market was shifting toward cloud-based solutions, albeit at different rates across different segments. Over time, it was assumed that almost all software would be deployed in the cloud, once latency issues had been resolved.
However, there are now signs that the direction of travel may become less one-way than previously assumed. A growing number of market participants appear to be reassessing on-premise deployments. It remains too early to determine the scale or durability of this shift and whether it represents a meaningful reversal rather than a temporary correction within the broader transition to cloud-based software. This insight explores some of the key factors driving renewed interest in on-premise solutions, and the potential implications this may have over time.
What’s led to the shift?
The shift toward on-premise deployment has largely been driven by the growth of AI, although there are several aspects to this. Firstly, since Anthropic teased its Mythos model and the many hundreds of thousands of vulnerabilities it detected while being trialed by leading cloud infrastructure providers, there seems to be growing concern around the security of cloud infrastructure. Whether this claim has any merit is outside the scope of this article, but the general sentiment has been one of caution.
Another factor is that some of the economic assumptions underpinning the ‘cloud-first’ model are coming under greater scrutiny. While on-demand infrastructure has long been associated with flexibility and lower upfront costs, this equation can look less compelling for AI-intensive workloads such as model training and large-scale inference. In these cases, the cost of renting high-performance compute in the cloud can be substantial. For organizations with consistently high utilization, outright ownership of hardware may offer a more attractive economic profile over time. Where infrastructure is being run continuously, the balance can begin to shift back in favor of on-premise deployment.
Regulatory considerations are also becoming more important to this discussion. As governments introduce new digital privacy, data sovereignty, and cyber resilience requirements, organizations are facing greater scrutiny of where data is stored, who controls access and encryption, and how systems are secured. In multi-tenant cloud environments, demonstrating compliance across these areas can be more complex, particularly in highly-regulated sectors. As a result, on-premise deployments may regain some appeal where greater control, auditability, and certainty around compliance are considered strategically important.
Will this trend continue?
Our view is that this shift is more likely to represent a correction than a sustained reversal. Some of the recent concerns around cloud infrastructure may prove to be overstated, particularly where they have been amplified by high-profile commentary such as Anthropic’s Mythos communications. In addition, much of the compute required for AI-enabled functionality in WMS is likely to sit with the software vendor rather than the warehouse operator or end-customer, unless the solution has been heavily customized or developed in-house.
Regulatory pressures are part of that reassessment, particularly as data sovereignty, cyber resilience, and auditability move higher up the agenda. This does not necessarily mean cloud deployment is structurally disadvantaged over the long term. In response, major infrastructure providers are evolving their offerings to address these concerns more directly, including via sovereign cloud models, tighter data residency controls, customer-managed encryption, and more isolated deployment options. Recent examples include AWS’s European Sovereign Cloud, which is designed as a physically and logically separate EU environment, and the broader move by hyperscalers toward sovereignty-focused architectures and local operational control.
The more important variable is likely to be the economics of cloud infrastructure. If demand for AI-related compute continues to grow faster than available capacity, cloud costs could rise, making the trade-off between cloud and on-premise deployments less straightforward than it has appeared in recent years. The key question will be whether the flexibility, scalability, and ease of deployment offered by cloud solutions continue to outweigh potentially higher hosting costs.
At a minimum, this trend underlines the extent to which AI is reshaping the software landscape, forcing CIOs and CTOs to reassess assumptions that previously appeared settled. For warehouse software buyers and vendors alike, the issue is no longer simply cloud versus on-premise, but also how infrastructure decisions should evolve in response to changing technical, economic, and regulatory realities. Our research into warehouse software will provide you with the clarity and evidence-based insights needed to help navigate the growing storm.
How can our research help you?
To learn more about how the warehouse software market is evolving and how it relates to the wider warehouse automation sector, please contact us to learn more about our research.
