Cloudera executives predict that in 2026, there will be radical changes in the way organizations build, manage, and store data for artificial intelligence, including short-lived AI-generated applications, increased oversight of “Frankenstein” systems, and pressure on data storage capacity.
Their predictions indicate a period of rapid experimentation. It also highlights the growing concerns about cost, control, and infrastructure limitations as AI becomes more deeply embedded in daily operations.
Chris Royles, Cloudera's field CTO for EMEA, says AI will change the way software is conceived and delivered.
“In 2026, AI will begin to fundamentally change the way we think about apps, how they function, and how they are built. Today, apps are declarative and have millions of lines of code to follow set rules. But AI breaks the rulebook and removes these constraints. With just a few lines of code and prompts, users will be able to request temporary specialized modules that replace the need for specialized apps, and AI essentially takes on the role of operating systems and app developers.
“For example, a user can ask the AI to coordinate a call between team members. Once the feature has accomplished its goal, it will exit. Then, in the background, the AI will silently refine the module based on its own assessment of experience and feedback until the same feature is needed again. These ‘single-use’ apps constantly evolve through self-learning and can be built and rebuilt in seconds.
“Strong security and governance are critical to this application development model, as organizations need visibility into the inference processes that AI uses to create and evolve each module. This ensures that errors are fixed and improvements are made securely and transparently.”
Royles expects this model to change the role of traditional long-lived applications. It may also change the way organizations think about software lifecycle and maintenance.
patchwork system
Cloudera executives also warn that early-stage AI builds can run into financial and compliance barriers. Many organizations combine various AI components into a single service.
Paul Mackay, RVP Cloud EMEA & APAC at Cloudera, said some of these systems are already under consideration.
“Many organizations will begin to shelve the 'Frankenstein' AI applications they built for specific business use cases as cost spirals and governance concerns grow. As organizations begin to deploy AI tools and applications, they run the risk of stitching together agent AI, large-scale language models, low-code and no-code platforms, and open source components in the name of innovation. As a result, even AI helpdesks and LLMs assembled from no-code forms become patchwork. But these patchwork systems can become money pits, caught up in unpredictable token usage and increased computing demands. They are also proving increasingly difficult to manage effectively.
“As observability and governance take center stage, organizations will reevaluate whether these applications need to be rebuilt or completely redesigned. Maintaining visibility into the data feeding these systems during development or rebuilding can help organizations regain control over compliance and costs. Without this, many ambitious AI projects risk becoming too expensive to sustain.”
Cloudera expects this reassessment to drive a more structured approach to AI design. It could also divert spending from experimental projects that lack clear controls.
autonomous monitoring
Along with concerns about ad-hoc builds, the company is looking forward to a new role for AI in governance itself. These exist within existing data operations.
“In 2026, we will see the emergence of specialized AI agents focused on data governance. These digital colleagues will continuously monitor, classify, and protect data wherever it resides, ensuring that governance becomes an always-on function built into daily operations,” said Wim Stoop, senior director at Cloudera.
“For example, compliance agents can flag risks across business systems, while security agents automatically adjust permissions as new data enters the environment, without human intervention. As these digital colleagues become part of the organizational structure, businesses will need to manage them as they would human teams, tracking their skills, performance, and collaboration through a new framework of “agent resource management.”
“Governance will no longer be something humans do, it will be something humans oversee. Rather than manually enforcing every rule, humans will ‘govern governance’ and shape processes as they run. As trust in these systems increases, we can step back and let continuous governance take over, delivering cleaner data, stronger compliance, and AI-enabled insights that drive real business value.”
Stoops said the transition will create a new management structure centered around AI agents. They will also need new tools to track their behavior.
data limit
Cloudera's outlook also includes pressure on data storage. The company sees a shift in the way organizations handle human-generated synthetic information.
“As global data storage approaches breaking point, the value of human-generated data will rise exponentially. The days of digital hoarding, holding everything just because storage was cheap, are over, and organizations will have to decide what is worth keeping and what needs to be produced again. AI-generated data will be disposable, created and updated on demand rather than being stored indefinitely.”
“This shift will redefine how data is valued. Companies will race to source and protect real, human-generated data to train and differentiate AI models. This shift will spark a new data economy that values originality, context, and quality over sheer volume, redefining the value of information in the age of synthetic content.”
Cloudera executives expect these trends to shape AI and data infrastructure investment decisions over the next year as companies test new applications and consider the long-term costs of their data strategies.
