IDC’s Global DataSphere Forecast estimates that global data volume will surge to 393.9 zettabytes by 2028. This growth puts business continuity, data protection, governance, and the cost of sustaining them under greater pressure. In Asia-Pacific, where rapid digitalisation, expanding cloud adoption, and data localisation requirements are accelerating data creation, organisations face an even steeper challenge in managing scale and complexity. This turns resilience into an AI economics issue, where every additional dataset retained and protected compounds spend across storage, backup operations, compliance overhead, and downstream AI quality and remediation.
Instead of focusing on whether organizations are backing up more data or adding more cloud security controls, the focus has shifted to whether those investments are improving business resilience in a way that is economically sustainable. Backup cannot be treated as an insurance policy that simply expands indefinitely. Without clear retention policies and strong governance, data resilience programs become financially draining, operationally burdensome, and harder to justify. The priority should be protecting the right data, at the right level, for the right recovery outcomes when disruption hits.
Governance is what makes resilience targeted, not indiscriminate
Understanding the data estate is where effective data resilience begins. According to Deloitte’s ‘The State of AI in the Enterprise 2026’ report, while 72 percent of organizations in Singapore plan to deploy agentic AI across operational areas within two years, only 14 percent report having a mature model for agentic AI governance, well below the global average of 21 percent. When AI agents make operational decisions autonomously, the cost of ungoverned data stops being a storage problem and starts being a business risk. Organizations need clarity on what data exists, how it is used, and what recovery expectations apply. Without that visibility, everything tends to be treated as equally critical, which quickly leads to oversized backup environments and unclear recovery priorities.
Governance provides the structure that allows organizations to prioritize protection. When datasets are classified according to business impact, protection levels can be tiered accordingly. Other datasets may be protected through lower-cost approaches or retained for shorter periods.
Determining what is critical is not purely a technical exercise. It depends on business commitments and the consequences of disruption, such as regulatory penalties, contractual obligations, operational risk, or reputational damage. When organizations frame resilience decisions through the lens of governance, protection, and retention become deliberate choices aligned to business risk rather than default IT configurations.
Stop paying twice for bad data
When governance priorities are unclear, many organizations default to keeping and backing up data “just in case.” Over time, that approach creates large volumes of information that provide little operational value.
IDC’s 2025 Survey Spotlight reveals that 64 percent of organizations are not highly effective at identifying, managing, and disposing of redundant, obsolete, and trivial (ROT) data. Yet organizations still store and protect this data, expanding backup sets and increasing recovery complexity. Larger backup environments require more data to be validated and restored before trusted operations can resume. The impact extends beyond infrastructure costs.
In organizations adopting AI-driven workflows, the consequences may be magnified. Poorly governed data often flows directly into analytics pipelines and AI models, introducing noise and reducing reliability in the insights. The result is a cycle where organizations “pay twice” by investing resources to store and protect low-value data, and then investing again to correct the problems that data creates in downstream systems.
Recovery testing proves governance decisions work in production
Governance strategies only matter if they produce reliable recovery outcomes. Regular restore and disaster recovery testing validates whether protection tiers actually match business priorities. These tests often surface critical insights, such as data that was backed up but did not contribute to recovery, or systems that require stronger protection than originally assumed. They can also expose hidden dependencies across data pipelines, where data lineage helps teams restore in the right order to resume trusted operations without restoring everything.
Leaders can track a small set of indicators to maintain focus. These include whether disaster recovery plans are tested regularly, whether recovery time objectives are clearly defined, whether tests consistently meet those objectives, and what improvements are implemented after each cycle.
Over time, this creates a feedback loop that strengthens governance. Insights from recovery testing inform cleanup efforts, policy updates, and operational improvements, helping organizations keep backup environments efficient and aligned with business needs.
Reducing data sprawl prevents resilience costs from compounding
In hybrid and multi-cloud environments, uncontrolled replication increases the volume of data that is secured, governed, and backed up.The IDC FutureScape: Worldwide Generative Artificial Intelligence 2025 Predictions suggests that 45 percent of Asia/Pacific organizations will adopt performance-intensive scale-out storage and unified data management by 2027 in order to accelerate AI and analytics. This adds to the total cost of AI adoption when data spreads across too many systems, copies, and unmanaged pathways.
That is why data movement into third-party SaaS platforms and external services should be treated as a deliberate governance decision, not a convenience. Once data leaves managed environments, visibility drops, controls become harder to enforce, and recovery becomes more difficult to coordinate.
Consistency across environments matters just as much. On-premises and cloud platforms need to be governed in the same manner to avoid managing and protecting data in fragmented ways that may encourage duplicate datasets and bloated backup environments. Open standards such as the Iceberg REST Catalog protocol can help by improving interoperability across engines and catalogs, reducing the need to create extra copies simply to make data usable across platforms.
The result is fewer duplicates, clearer ownership and retention, and a smaller, cleaner backup footprint that is easier to govern and manage, and costs less to maintain.
What leaders should take away
Resilience in modern enterprises is not about creating more copies of everything or piling on controls. Instead, it is about making intentional, governed decisions so organizations avoid carrying an ever-expanding bill for data that is redundant, obsolete, trivial, or simply unknown.
Governance is the mechanism that optimizes backup spend, shortens recovery, and improves AI reliability. With this, companies will stop paying premium prices to protect data they did not understand, did not need, or should not have kept in the first place.
Carolyn Duby is Field CTO and Cyber Security GTM Lead at Cloudera.
Carolyn is a cybersecurity, data, and AI expert with 30+ years of experience across various industries, from high-performance medical devices to cybersecurity and distributed big data solutions. She brings expertise in helping customers build secure, efficient AI applications that enhance brand reputation, strengthen cyber defenses, and shape data strategy.
Prior to joining Cloudera in 2021, she held senior positions at Pathfinder Solutions, Dell Secureworks, and Hortonworks. She joined Cloudera, as Principal Solutions Engineer and now serves as Field CTO and Cybersecurity AI Strategist.
TNGlobal INSIDER publishes contributions relevant to entrepreneurship and innovation. You may submit your own original or published contributions subject to editorial discretion.
Featured image: geralt on Pixabay
Why fuel shocks are turning EV adoption into a policy and fiscal issue in Asia