IBM, targeting the enterprise AI "operation phase"… The key to victory lies in hybrid cloud and governance

IBM is continuously enhancing its presence in the enterprise AI market by leveraging the three core advantages of “speed,” “cost reduction,” and “security.” Its key lies in the “Hybrid AI” strategy, which enables AI to operate based on actual workflows even in highly regulated and complex enterprise environments.

IBM describes this approach as a model of “workload prioritization,” “ubiquitous deployment,” “centralized governance,” and “infrastructure abstraction.” In short, this means helping enterprises introduce AI in a controlled manner while avoiding dependence on specific cloud services or single systems. IBM has also simultaneously implemented a “Customer Zero” strategy, which involves first applying AI products within its own internal organization—spanning 175 countries and employing 280k staff—before rolling out to external clients. This approach allows for internal validation prior to market deployment.

Market response has been equally positive. IBM announced revenue growth in its software and infrastructure divisions in its mid-April earnings report. Notably, revenue from its next-generation mainframe product line surged by 48%. This is seen as a result of the combined demand for AI and existing enterprise infrastructure.

Moving from AI experimentation to operational deployment

At the upcoming “IBM Think” event on May 12, how enterprises transition from AI experimentation to actual operational environments is expected to be a major focus. Among the key topics are “Agent AI,” which manages multiple AI agents simultaneously, and the governance systems used to unify control over these agents.

In this process, IBM is positioning itself as the “control layer” for enterprise AI. Its vision includes providing hybrid cloud integration, trusted data pipelines, and multi-agent operational infrastructure. This can be interpreted as a strategy that emphasizes not competing directly with large language models (LLMs), but rather helping enterprises integrate AI stably into real-world business operations.

Earlier this year, IBM launched “IBM Sovereign Core,” which reflects the same philosophy. The platform aims to give enterprises and government agencies more direct control over AI and cloud workloads. In the context of increasing data sovereignty and compliance requirements, IBM is attempting to enhance flexibility through open frameworks and a partner ecosystem.

Quantum computing and “Post-Quantum Security” preparations

Another major pillar for IBM is quantum computing. At each year’s IBM Think event, new developments related to quantum technology are unveiled, and this year is likely to include a related roadmap.

A typical example is IBM’s collaboration with Cisco to advance the so-called “Quantum Internet.” The goal is to connect quantum computers over long distances and plan to expand the network into a distributed architecture comprising dozens of devices.

Meanwhile, security measures are also progressing in parallel. IBM is preparing for “Post-Quantum Security” to address the potential phasing out of current public key cryptography algorithms by 2035. This is because widely used encryption methods today may become vulnerable in the face of future quantum computers.

An IBM security executive emphasized in a recent interview that what is needed now is “cryptographic agility,” meaning the ability to rapidly switch between cryptographic systems. The explanation is that if systems continue to rely on fixed encryption structures as in the past, they will struggle to respond to emerging threats.

Collaborating with NVIDIA and Arm to enhance data accessibility

Beyond its own technology, IBM is strengthening its enterprise AI strategy through partnerships and acquisitions. In March, IBM announced an expansion of its collaboration with NVIDIA ($NVDA) to support large-scale AI deployment. Specific measures include connecting IBM tools with NVIDIA’s NeMo Retriever to accelerate document extraction, and integrating IBM’s unified data access layer with NVIDIA’s GPU pipelines.

Earlier this month, IBM partnered with Arm to announce a new dual-architecture hardware plan for AI and data-intensive workloads. This initiative aims to meet enterprise needs for flexible deployment across multiple system environments without relying on a specific semiconductor architecture.

Additionally, IBM acquired streaming data company Confluent in December last year, enhancing its real-time data processing capabilities. This move is interpreted as addressing enterprise demands for instant access to reliable data within complex hybrid cloud environments.

Focusing on “Trusted AI Operations”

IBM’s direction is very clear. It is not merely chasing cutting-edge models but focusing on operating trustworthy AI within actual enterprise environments. This aligns with the reality of the large enterprise market—compared to flashy technological demonstrations, post-deployment management, compliance, security, and data integration are more critical.

The key question is whether IBM can become the “record system” and core operational platform for enterprise AI. Will it remain one of many vendors in the AI tech stack, or can it become the central infrastructure for enterprise AI? The answer is expected to become clearer around the IBM Think conference.

IBM’s simultaneous initiatives in hybrid AI, quantum computing, and data access strategies indicate that the enterprise AI market has shifted from the “experiment” phase to the “operational” competition stage.

TP AI Notice: This article is summarized based on the language model of TokenPost.ai. It may omit main content or differ from actual facts.