Lenovo outlines a hybrid AI approach at CES 2026

• At CES 2026, Lenovo outlined a move toward system-level AI across devices.
• Its hybrid approach blends on-device and cloud AI for more contextual, less intrusive use.

AI is becoming less about standalone tools and more about how intelligence is woven into everyday systems. At CES 2026 in Las Vegas, Lenovo used its Tech World event to outline how it sees that shift taking shape across personal devices and enterprise environments, with AI operating more quietly in the background rather than as something users must actively manage.

The company focused on what it calls “Hybrid AI,” an approach that combines on-device processing with cloud-based models. Instead of presenting AI as a single product or feature, Lenovo framed it as a layer that connects devices, applications, and services. The goal, as described during the event, is to make AI more responsive to context while keeping tighter control over where data is processed and stored.

At the centre of that strategy is Lenovo Qira, a system-level AI layer designed to work across multiple devices. Qira is meant to carry context with users as they move between PCs, smartphones, tablets, and other form factors, rather than forcing them to start over with each device. Lenovo is extending the same system to Motorola devices, signalling an effort to create a shared intelligence layer across its broader hardware portfolio.

Dan Dery, vice-president of AI Ecosystem in Lenovo’s Intelligent Devices Group, said the company is trying to move beyond the idea of AI as a tool that users open and close. He explained that Qira is intended to feel like an intelligence that works alongside people across their devices, adapting over time rather than responding only when prompted.

From assistants to system-level intelligence

Rather than requiring users to switch apps or issue explicit commands, Qira is designed to remain available at the system level. It can be activated by voice, a dedicated key, or on-screen controls, but it can also stay in the background until needed. Over time, it builds an understanding of user preferences based on interactions and content that users choose to share, with the aim of offering support at the right moment rather than interrupting workflows.

Lenovo described this shift as a move away from app-based AI toward ambient intelligence. In practice, that means AI that can suggest next steps, assist with writing, or summarise activity without forcing users to break focus. The company highlighted several core experiences built around this idea, including tools that help users catch up after stepping away, support live interaction during screen or camera sharing, and provide real-time transcription or translation during meetings when enabled.

For writing tasks, Qira can assist directly within documents or messages instead of pushing users into separate tools. Lenovo said the system adapts to tone and intent, helping users shape drafts or refine ideas while staying on the same canvas. In collaborative settings, Qira can stay active during calls or presentations, responding to both spoken input and visual context.

Underpinning these experiences are three core capabilities Lenovo emphasised: presence, action, and perception. Presence refers to how Qira remains accessible across devices in a consistent way. Action relates to its ability to carry out tasks locally on a device, including offline, without relying entirely on cloud services. Perception describes how Qira builds a knowledge base from user-approved data, documents, and interactions, while keeping privacy controls in place.

Privacy, control, and enterprise implications

Privacy was a recurring theme throughout Lenovo’s presentation. The company said its hybrid approach prioritises on-device processing to keep personal data local whenever possible, while secure cloud services extend capability when needed. Lenovo framed this as privacy by design, with user consent and control treated as foundational rather than optional.

To support this model, Lenovo is working with a range of partners. Microsoft’s AI stack, through Windows Foundry and Azure, underpins coordination across Windows devices. Stability AI supports on-device image generation within Creator Zone on PCs. Other integrations cover note-taking, research, and travel planning, all designed to surface relevant information without forcing users to leave their current task.

While much of the discussion centred on personal devices, Lenovo also linked these ideas to enterprise use. The company argued that the same hybrid AI approach can help organisations apply AI to their own data and processes, rather than relying only on generic models.

Lenovo chairman and chief executive Yuanqing Yang said AI is beginning to draw more directly from individual habits, language, and experiences, which he believes will help people be more creative and intuitive in their work. For businesses, he said the shift goes further, with AI moving beyond workflow optimisation and into systems that can learn from proprietary data and apply an organisation’s own decision logic over time.

Taken together, Lenovo’s CES presentation positioned AI as less visible but more embedded, with intelligence spread across devices and environments rather than concentrated in a single interface. How well that approach works will depend on execution, but the direction is clear: AI is being designed to fit into existing workflows, not replace them.

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology eventsclick here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.