2. Cloud backups
The second precedence can also be unsurprising given current information. Of 284 enterprises who commented, 173 mentioned that they wanted a technique to again up cloud elements within the occasion of cloud outages. This, they are saying, is much more difficult than senior administration thinks. First, you must resolve simply what actually wants backed up. “You possibly can’t completely immunize your self in opposition to an enormous cloud or Web downside,” say planners. Most cloud outages, they word, resolve in a most of some hours, so you’ll be able to let some purposes experience issues out. When you already know the “what,” you’ll be able to have a look at the “how.” Is multi-cloud the most effective strategy, or are you able to construct out some capability within the information heart? Enterprises word that constructing in resilience in any type might require redesigning some purposes to make cloud-hosted parts transportable, and that may additionally imply the place utility information is saved and the way entry to it’s linked.
3. Infrastructure simplification
Precedence three is managing the technical complexity of infrastructure, cited by 139 enterprises. “We’ve got too many issues to purchase and to handle,” one planner mentioned. “Too many sources, too many applied sciences.” No person thinks they’ll do some huge fork-lift restructuring (there’s no finances), however they do consider that present tasks could be aligned to a long-term simplification technique. This, apparently, is seen by over 100 of the group as lowering the variety of distributors. They suppose that “lock-in” is a small value to pay for higher effectivity and discount in operations complexity, integration, and fault isolation. That is the largest shift in opposition to multi-vendor or open infrastructure I’ve ever seen.
4. Prioritize governance
The shut quantity 4 precedence is extra administrative than infrastructure-related; 124 enterprises in our group mentioned they wanted to “completely revamp governance.” Sure, AI is an enormous issue on this, however so is the elastic-hosting mannequin of cloud and multi-cloud, and the “sovereignty” points related to working throughout a number of jurisdictions, and the more and more chaotic nature of laws. The proportion of enterprises who say they want some formal “authorities affairs” enter to administration practices has elevated from 12% in 2020 to 47% for 2026. For instance, EU cloud and AI sovereignty issues influence plans for each AI and cloud utility resilience.
The most important downside, these enterprises say, is that governance has tended to be utilized to tasks on the planning degree, that means that absent main tasks, governance tended to limp alongside primarily based on growing old evaluations. Enterprises word that, like AI, orderly expansions in how purposes and information are used can introduce governance points, similar to modifications in legal guidelines and laws. AI complicates this as a result of it’s tough or unattainable to know simply what information AI is accessing, if there aren’t any filters on information availability. All this can be a governance problem, however it could actually pale compared to the truth that corporations aren’t used to even enthusiastic about governance absent a challenge framework. Do that you must create “governance tasks?” If that’s the case, how are they justified, funded? The place there are exhausting modifications in legislation or laws, there are procedures, however not a lot with different challenges. AI brokers, even workflow brokers, can creep into governance issues as utilization grows, for instance.
5. Value administration
The ultimate precedence on our listing, with 108 enterprises citing it, is in some ways a barrier to fulfilling any of the opposite targets they determine: Do extra for much less. Of our 284 enterprises, 226 mentioned that they have been beneath extra finances strain for 2026, and solely 9 mentioned that they had much less strain (for the remaining, pressures have been the identical). It’s fascinating, although, that quantity 5 on the priorities listing is the bottom scored for price administration since 2008/2009.
The fascinating factor about this explicit precedence is that, not like prior years the place the “price” being managed was presumed to be the capital price of the applied sciences concerned, the gear and software program, the main target for 2026 is the overall price, what can be categorised as “complete price of possession” or TCO. This may be simple, or at the least attainable, within the context of conventional challenge pondering, however so many of those priorities blur the traces between “tasks” that require assessment, justification, and approval, and regular day-to-day enterprise choices that normally dodge a lot of that formality. How do you assess the TCO of AI effectivity optimization total, or cloud utility resilience, or governance?
