AI diffusion shifts safety focus... Companies surpass "backup" to compete in "AI resilience"

As artificial intelligence rapidly penetrates frontline enterprise operations, the focus of the security industry is also changing. Today, it’s no longer just about simple data backups, but about how to “trust” the complex operational environments created by AI and whether it can recover quickly from issues, which is becoming a core competitive advantage.

Backup and disaster recovery specialist Veeam recently highlighted this shift at the “VeeamON” event. Veeam CEO Anand Eswaran outlined three key reasons why establishing a foundation of trust in AI infrastructure has become important. He pointed out that enterprise adoption of AI agents has been broader than expected; investments in AI infrastructure are projected to reach $3 trillion by 2028, approximately 4,496 trillion Korean won; however, compared to this, the trust infrastructure supporting all of this has not yet been fully established.

Industry analyst Dave Vellante noted, “Impressively, 81% of companies are already running agents.” “Nearly half of them may not even be fully aware of this.” He further commented, “An environment with 82 agents per person is coming, and AI is actually eroding the original boundaries of trust.”

Veeam, Expanding from Backup to an “AI Trust Layer” Service Provider

The biggest change revealed at this event is that Veeam is no longer content with being just a traditional backup company. The company is expanding its business scope based on data security and recovery capabilities to become a comprehensive platform provider responsible for “AI resilience.”

Veeam is especially focused on building a “trust layer,” integrating the sensitivity, access permissions, compliance, and privacy protections of data used by AI into a unified system. As a result of this strategy, the company acquired a data security firm and launched the “Veeam DataAI Command Platform,” which combines data security, governance, compliance, privacy, and recovery.

This aligns with the core issue of the AI era: it’s not just about model performance, but about “what has been learned, what data has been accessed, and how to recover when problems occur.” In other words, AI resilience is being redefined: it is no longer a sub-concept of cybersecurity but a foundational infrastructure supporting overall enterprise operations.

AI Breaks Department Silos, Security and Governance Face Integration Pressure

Veeam believes that AI is breaking down internal “silos” within enterprises, i.e., the fragmented structure of departmental data and systems. In the past, security, operations, and data management could operate independently, but because AI requires cross-organizational data operations, this separation is becoming increasingly untenable.

As a result, companies can no longer view AI and security separately. AI results generated from untrusted data can amplify decision errors, and uncontrolled AI can widen security vulnerabilities. This means that the key to AI application success or failure has shifted from “how to effectively use AI” to “how trustworthy AI is.”

ZK Research’s Zeus Kravala pointed out that although AI promises the “democratization” of expertise, in reality, users still need domain knowledge to verify results. Because it’s difficult to fully trust the answers AI provides. This is precisely why an AI trust layer is needed.

Network Resilience Is Also Evolving Toward AI-Driven Approaches

In related ecosystems, collaboration continues to expand. Infrastructure companies like ExaGrid and Everpure are partnering with Veeam to strengthen AI-based anomaly detection and backup environment protection functions. They analyze backup data behavior patterns through neural networks to automatically detect suspicious activities and alert administrators.

This indicates that backup systems are no longer just repositories but have become the “last line of defense” to protect enterprises during attacks. Especially amid the accelerating threats of ransomware and data theft, ensuring recoverability and data integrity has become more strategically important than ever.

Attackers are also leveraging AI to increase the speed and scale of threats. Ray Umeri of Coveware, a Veeam subsidiary, analyzed that while traditional ransomware strategies haven’t changed much, AI makes them faster and more sophisticated. Lowering the attack threshold significantly increases the complexity for enterprises to respond.

As AI investments expand, companies should focus on “control” rather than just “speed”

Enterprise IT departments face pressure to improve efficiency amid expanding AI infrastructure investments and limited budgets. Therefore, the shift from primarily building in-house equipment to cloud-based operational models is accelerating. However, whether data is stored locally or in the cloud, ultimate responsibility for protection still lies with the client company, and this remains unchanged.

Veeam and industry experts agree that although many enterprises are rapidly expanding AI applications, they often lack a full understanding of where their data resides and what risks it is exposed to. Evidence collection needed for audits and compliance is also often insufficient.

Ultimately, the message from this VeeamON is clear: in the AI era, competitiveness no longer depends solely on application speed, but on trustworthy data, recoverable systems, and a unified “AI resilience” that manages all of these. As AI increasingly becomes the core of enterprise operations, trust infrastructure is no longer optional but a necessary element.

TP AI Notes: This article is summarized based on the TokenPost.ai language model. Some key content may be omitted or may differ from actual facts.