BlackBerry's Best Quarter in Years Was Powered by Physical AI

Every earnings season, executives explain why AI is not a threat to their business. Most of those explanations involve retraining, productivity gains, or adjacent opportunities.

BlackBerry CEO John Giamatteo offered a different argument on the company's Q4 FY2026 earnings call in April, and it is more structurally compelling than most.

"AI is probabilistic by nature, meaning outputs can vary," Giamatteo said on the call. "But the QNX platform is deterministic, delivering the same result every time without exception. That distinction matters enormously when our software controls the vehicle safety features such as adaptive cruise control or autonomous drive."

The logic extends beyond automotive. An AI-driven heart pump cannot have a probabilistic output. An autonomous surgical robot cannot produce a result that varies. An industrial automation system controlling high-voltage equipment cannot fail unpredictably.

In every environment where AI is being deployed to take autonomous action in the physical world, there is a layer of software beneath the model that has to perform exactly as specified, every single time, without exception.

That layer is what QNX does. And as AI moves from digital interfaces into physical systems, the demand for that layer is accelerating.

The Physical AI Thesis

Giamatteo framed physical AI, autonomous robots, intelligent edge devices, and industrial automation, as one of QNX's most significant growth accelerators on the call. Design wins across robotics, industrial automation, and medical devices are already converting into backlog.

In Q4 FY2026 alone, QNX secured a significant win for its General Embedded Development Platform in industrial automation controls for a major North American OEM.

Johnson & Johnson selected QNX OS for Safety to power a new AI-driven heart pump, one of the most specific examples of QNX serving as the operating foundation for an AI system in a life-critical medical device.

The medical instrumentation vertical, Giamatteo said, is showing a rapidly increasing contribution alongside industrial automation and robotics.

The automotive blueprint is the template for all of it. QNX already runs in the most complex consumer device ever built, a car, where it powers adaptive cruise control, autonomous driving systems, and digital cockpit infrastructure.

The car is, as Giamatteo put it, "essentially a robot on wheels." The credibility QNX has earned in automotive including the safety certifications, the decades of flawless execution, and the ISO 26262 functional safety standard, translates directly into the robotics and physical AI environments that are now emerging.

ARM validated that argument publicly. At the launch event for its ARM AGI CPU for physical AI, CEO Rene Haas named QNX as one of the foundational software ecosystem partners supporting ARM's aspirations in the space. NVIDIA and Qualcomm, both pushing aggressively into physical AI, are also established QNX partners.

The Backlog That Proves It

QNX's royalty backlog grew to approximately $950 million in FY2026. The backlog grew more in FY2026 than BlackBerry recognized in its P&L during the year, meaning future revenue generation is accelerating faster than current recognition.

That pattern, adding backlog above the rate of P&L recognition, is the clearest possible leading indicator of sustained revenue growth.

QNX delivered record revenue of $78.7 million in Q4 FY2026, 20% year-over-year growth, and achieved Rule of 40 for both the quarter and the full fiscal year, 14% annual growth and $71 million of adjusted EBITDA at a 26% margin. QNX fiscal year revenue guidance for FY2027 targets up to 15% growth to $290 million to $307 million.

The General Embedded Market now represents approximately 20% of QNX revenue and is growing faster than automotive. Giamatteo described robotics as "probably going to be one of the faster-growing segments" of the three GEM verticals the company is targeting.

"Far from being complacent, we see AI as a net tailwind for our business rather than a threat," Giamatteo said. The argument is specific enough to be credible.

QNX is not competing with AI models. It is the platform those models run on when they take physical action in regulated, safety-critical environments, and the bar for replacing a certified, proven, deterministic operating system in those environments is, as Giamatteo put it, a cost-benefit calculation that "no responsible OEM will make."

BlackBerry declared its turnaround complete on the April 9 call. The growth story that follows is built on a bet that physical AI will be the defining enterprise AI story of the next decade.