Artificial intelligence is moving into a new chapter where experimentation gives way to implementation, and surrounding infrastructure matters as much as the models themselves. In a recent IBM Think article, contributors shared how they expect AI and related technologies to evolve in 2026.
Their outlook reveals how hardware and model governance are being reevaluated in response to AI’s use cases.
Growing interconnected environments have shifted the AI conversation.
Topics like agent orchestration and hardware constraints are now central concerns for people deploying AI in real-world settings. This has led decision makers to extend their focus past what AI can do by looking at how it should be developed to make it genuinely useful and sustainable.
Why It Matters: This moment in AI development presents new questions for organizations that want to adopt and deploy intelligent systems responsibly. Models need to operate within environments that support transparency, security, and flexible design. These requirements are beginning to shape technical architecture and how AI tools are funded and evaluated.
- Quantum Computing is Beginning to Support Real-World Experimentation: Developers are starting to use quantum systems alongside traditional compute infrastructure to explore problems that benefit from new approaches. Some of these efforts are focused on molecular simulation and optimization in financial services. Others involve building hybrid architectures that connect quantum processors with classical CPUs and newer forms of AI-accelerating hardware. Quantum coding assistants are being introduced to help write more efficient programs for this environment, lowering the barrier for developers who are entering the field.
- AI Development is Focusing More on Orchestration and Systems-Level Design: Instead of relying on a single large model to solve a task, developers are assembling systems that combine models with tools and data pipelines. These systems are designed to handle processes that unfold over time. This allows for greater flexibility and enables adjustments to how tasks are completed based on context. Tools are being built to manage interactions between these components, creating more durable and customizable workflows. The end goal is to produce environments where different AI elements contribute to a larger process.
- Agents are Taking on Roles Spanning Multiple Applications and Contexts: Agents are not limited to responding to prompts or completing simple actions. They are being developed to carry out extended sequences of tasks and operate across different software environments. In many cases, these agents can initiate their own steps once a goal has been defined. Developers are experimenting with interfaces that let users assign tasks and review progress across a set of agents, which creates a more coordinated and user-directed experience.
- Open-Source Models Are Becoming More Specialized and Easier to Manage: Rather than training extremely large models for general use, developers are focusing on smaller models that are tuned for specific domains or applications. Models are being built to run on local devices or more modest infrastructure. Their design is often shaped by operational requirements such as faster response times, lower energy use, or more reliable access to data. At the same time, open-source communities are working on governance practices that allow contributors to understand how models were trained and what data was used. This creates more clarity around model behavior and makes it easier to evaluate whether a system meets the requirements of a given use case.
- Visibility and Oversight are Becoming Necessary Parts of AI Deployment: Agents and automated systems becoming more active within business environments raise the need for tools that can monitor action alignment with policy or expectations. Some teams are developing identity systems that treat agents as unique entities within an access framework. These systems are being built to show which resources an agent has touched, what decisions it made, and how it arrived at those outcomes. For many organizations, this is becoming a requirement rather than a feature, especially when AI tools operate in environments where compliance or risk reduction are priorities.
Go Deeper -> The trends that will shape AI and tech in 2026 – IBM
Trusted insights for technology leaders
Our readers are CIOs, CTOs, and senior IT executives who rely on The National CIO Review for smart, curated takes on the trends shaping the enterprise, from GenAI to cybersecurity and beyond.
Subscribe to our 4x a week newsletter to keep up with the insights that matter.
