BlackBerry Expands QNX Footprint into Physical AI and Autonomous Systems
BlackBerry Ltd is expanding the application of its QNX operating system beyond traditional automotive sectors into the emerging field of “Physical AI.” According to company reports, the firm is seeing increased demand for QNX in robotics, autonomous vehicles, and industrial systems.
The company confirmed that Uber has selected QNX for its next generation of vehicles. Additionally, Momenta and XHEART have adopted the QNX OS for Safety to power their autonomous-driving platforms. BlackBerry noted that more than 20 companies are currently engaging with the company regarding NVIDIA-based platforms, with projects spanning humanoid robots, surgical robots, autonomous tractors, and drones.
While management characterizes Physical AI as an early-stage opportunity, the company expects its NVIDIA partnership, alongside the QNX Everywhere and QNX Launchpad initiatives, to drive future pipeline growth.
Financial Outlook and Market Sentiment
Despite the expansion of its technological footprint, BlackBerry has faced scrutiny regarding its share price performance. While revenue has increased, some market observers have questioned why the stock price has faced downward pressure despite a positive revenue outlook and new partnerships.
In recent earnings discussions, management indicated a cautious approach to growth, citing ongoing geopolitical tensions as a factor. Some analysts suggest that the company’s focus on secure communication may provide a strategic advantage in a volatile geopolitical climate, similar to the strategic importance of essential commodities.
Technical Developments in Local AI Integration
In a separate development circulating on social media platforms, unverified reports have surfaced regarding the integration of small-scale local AI models into Advanced Driver Assistance Systems (ADAS).
According to discussions on Reddit, a 1-billion parameter (1B) local AI model has been demonstrated in a driving assistant context. While some users have expressed skepticism regarding the reliability of a model of this size for critical driving tasks—noting that at such low quantization, the model may function more as a regression model than a sophisticated reasoning engine—others have defended the implementation. Specifically, some users noted that the technology being discussed may be a “Comma 4” system rather than a standard local Llama setup.
These reports remain unverified by independent technical audits, and the practical safety implications of relying on low-parameter local models for real-time driving decisions remain a subject of debate among developers and enthusiasts.