Customers
Kyndryl and Incore Bank collaborate with Google Cloud to help advance agentic AI customer onboarding
Press release
31 Aug 2026
At VMware Explore 2026, one message came through clearly: private cloud is becoming much more than a platform for traditional infrastructure. As organizations put artificial intelligence into production, they need private clouds that can support modern applications, agentic workflows and GPU-driven architectures while providing greater control over data, security and cost.
We spoke with Zeb Ahmed, Global Vice President, Cloud Offerings and Services at Kyndryl, about what he heard at the conference and how organizations can evolve their existing environments into platforms designed for the AI era.
What stood out to you at VMware Explore?
Zeb Ahmed: The conversation has moved beyond infrastructure virtualization. Private cloud is increasingly becoming a platform for containers, modern applications, AI and agentic operations.
That shift reflects what customers are working through today. They want to bring AI closer to critical enterprise data, manage cost more predictably and strengthen control over sensitive workloads – but they also expect the speed, automation and developer experience associated with public cloud. The challenge is to bring those capabilities together so organizations can place each workload where it makes the most business, operational and regulatory sense.
What makes a private cloud ready for the AI era?
Ahmed: An AI-ready private cloud needs to support the complete lifecycle of AI and agentic workloads. That includes Kubernetes-native platforms, access to appropriate compute, automated provisioning, observability, cost management and controls governing how models and agents interact with enterprise systems and data.
Data is especially important. Much of an organization’s critical data may remain on-premises. Rather than moving entire data estates before pursuing AI, organizations need a governed way to give models and agents access to the specific data required for each use case. Governance cannot be added after agents are deployed. It must be designed into the platform and operating model.
Platform engineering is the bridge between more traditional private clouds, which often rely on tickets, manual orchestration and fragmented automation, and the more dynamic next-generation private clouds. It gives developers standardized, reusable paths to build and deploy, while embedding identity, security and policy controls into the experience from the beginning.
Organizations are beginning to think beyond deploying individual agents toward governing entire populations of agents. They need to manage identities, access, model routing and approved actions while maintaining visibility and human oversight.
How is Kyndryl helping organizations make that transition?
Ahmed: We begin with the customer’s business priorities, workloads, existing environment and risk requirements. We then help assess the estate and determine what should be optimized, upgraded, rebuilt or integrated into a broader hybrid model.
That’s an important distinction. A software upgrade may address an immediate requirement, but modernization should also improve automation, developer experience, governance, resilience and the way the environment is operated.
Kyndryl Private Cloud brings those capabilities together in an integrated, scalable solution backed by agentic platform engineering. It delivers enterprise-grade performance and governance with a public-cloud-like experience, using predefined or custom-built architectures and hosted, dedicated or tailored deployment models.
Kyndryl supports that platform across the lifecycle, from advisory and architecture through migration and managed operations. We use automation and agentic analysis to help customers understand dependencies and assess the cost, value and risk of different modernization paths, so they can build an evidence-based business case before making major infrastructure decisions.
Our differentiation is that we carry those improvements from design through ongoing operations. We modernize with the realities of mission-critical environments in mind and help customers continue optimizing after the initial implementation.
Our expanded alliance with Broadcom strengthens that approach. VMware Cloud Foundation provides a common platform for virtualized and containerized workloads, while Kyndryl brings the expertise to design, modernize, integrate and operate it across complex environments.
What did you hear from customers about security and agentic AI?
Ahmed: One of the strongest themes we heard from customers was the need for AI assurance. Organizations are beginning to think beyond deploying individual agents toward governing entire populations of agents. They need to manage identities, access, model routing and approved actions while maintaining visibility and human oversight.
Kyndryl’s Agentic AI Framework and policy-as-code capabilities help keep agent actions within defined business and regulatory guardrails. Security also must extend through applications, infrastructure and recovery. Building those controls into the platform gives developers a simpler experience while helping the organization maintain a stronger security and compliance posture.
What should organizations prioritize now?
Ahmed: A software or infrastructure upgrade is not the same as modernization. An upgrade can address an immediate technology requirement, but continuous, effective modernization also improves automation, developer experience, governance, security and the operating model.
Organizations should start with the workloads and outcomes that matter most. Determine what data those workloads require, where they should run and what security, governance, performance and cost requirements must be met. Then they should assess whether a private cloud can support those needs as a platform, securely and economically.
Kyndryl works with customers to make that transition progressively, meeting them where they are to help turn private cloud into a secure, governed and continuously modernized foundation for AI.
