Qira, Lenovo’s largest AI initiative to date, is designed to function across Lenovo laptops and Motorola smartphones.
Qira, Lenovo’s largest AI initiative to date, is designed to function across Lenovo laptops and Motorola smartphones.
While the majority of interest in the AI competition is directed at model creators and cloud systems, Lenovo is in closer proximity to millions of users than many other firms. As the world’s leading PC manufacturer by unit volume, Lenovo distributes tens of millions of units annually. Its choices on what to send, package, and incorporate can directly influence how AI integrates into numerous daily experiences.
This context made Lenovo’s announcement at CES today significant. During a vibrant event on Tuesday at The Sphere in Las Vegas, it unveiled Qira, a system-level, cross-device AI assistant intended to operate seamlessly on Lenovo laptops and Motorola smartphones. It marks Lenovo’s most ambitious AI venture to date and provides a rare glimpse into how a major hardware company with global influence plans to deepen AI integration.
Jeff Snow, Lenovo’s AI product leader, explained how Qira developed, why the company is intentionally steering clear of a singular exclusive AI partnership, and what insights he gained from previous projects like Moto AI and Microsoft’s Recall fiasco.
Qira resulted from a discreet yet impactful internal restructuring less than a year ago, according to Snow. Lenovo consolidated AI teams from separate hardware divisions such as PCs, tablets, and smartphones into a centralized software-oriented unit that collaborates across the entire organization.
For an enterprise long focused on hardware SKUs and supply chains, this shift indicates a transition toward emphasizing AI in a prominent manner. “Our aim was to create a built-in cross-device intelligence that collaborates with you throughout the day, learns from your interactions, and can act on your behalf,” Snow articulated. He noted utilizing Qira’s on-device model during his CES flight to refine how to present information in meetings based on notes and documents on his laptop.
Qira is not centered around one premier AI model. Rather, it is modular. Beneath the surface, it blends local, on-device models with those based in the cloud, leveraging Microsoft and OpenAI infrastructure via Azure. The inclusion of Stability AI’s diffusion model is also present, in addition to partnerships with app-specific collaborators such as Notion and Perplexity.
“We sought to avoid binding ourselves to a single model,” Snow expressed. “This domain is evolving rapidly. Various tasks require differing tradeoffs concerning performance, quality, and cost.”
This approach contrasts with the trend from major AI laboratories, many of which would be eager to become the exclusive intelligence layer for a company with Lenovo’s scale. Lenovo believes that flexibility is crucial, particularly considering its oversight of one of the largest consumer computing distribution networks globally.
Snow previously contributed to Moto AI, Motorola’s assistant, which he noted had high initial engagement. Over half of Motorola users experimented with it, but retention rates were lacking. He stated that much of the experience resembled prompt-based chat functions available through other platforms.
“This experience led us to steer away from competing with chatbots,” Snow remarked. “Qira focuses on functionalities that chatbots cannot accomplish, such as continuity, context, and the ability to act directly on your device.”
Lenovo has also closely monitored the criticism surrounding Microsoft’s Recall feature. Snow indicated that Qira is intentionally designed with opt-in memory, consistent indicators, and transparent user controls from the beginning. Context ingestion is optional, recording is transparent, and nothing is collected without notice.
Cost pressures are a significant concern in this context. Memory expenses are increasing as AI demand challenges supply chains, and analysts anticipate that PC prices will rise in tandem. Snow mentioned that Qira does not elevate the minimum system requirements for PCs but performs optimally on higher-end machines with greater RAM. Lenovo is working to reduce the memory footprint of local models to 16 gigabytes of RAM while maintaining quality.
From a strategic perspective, Lenovo views Qira as both a strategy for customer retention and a safeguard against hardware commoditization. In the short term, it aims to encourage customers to remain within the Lenovo ecosystem through tighter integration of laptops and phones. In the longer run, Snow presented Qira as a means to differentiate Lenovo devices when specifications alone are insufficient.
