BlackBerry Bets on Automotive and Robotics Software After Losing Mobile Phones
CNBC's "The Tech Download" focuses on BlackBerry: after losing the smartphone war, the company is betting on automotive and robotics software. CEO John Giamatteo explains that the company has restructured around the real-time operating system QNX for software-defined vehicles and is focusing on physical artificial intelligence.
He describes QNX as "the operating system for cars." Of the approximately 90 million vehicles produced globally each year, about 25% use advanced software such as digital cockpits and advanced driver assistance systems. Company data indicates that QNX has been embedded in over 275 million vehicles. There are backlogged orders exceeding $950 million, most of which are still in the automotive sector, with physical artificial intelligence being the fastest-growing segment.
He assesses humanoid robots as being "probably still in the first game." Applications also include automated mobile robots for warehousing, surgical systems, and industrial automation. The selling point of QNX is determinism and safety certification, rather than probabilistic large models; he expressed caution about how much generative artificial intelligence to incorporate into product code, as failure is unacceptable.
The company has been profitable for several consecutive quarters, a rarity since the peak of its mobile phone era. After a quarterly report exceeded expectations in 2026, the stock price surged about 23% in one day. Chip collaborations include NVIDIA and AMD; in China, there are partnerships with autonomous driving companies like WeRide. The second growth point also includes general embedded systems and Alloy Core, with management stating they expect to secure their first design import in the current fiscal year.
Buyers are automotive and robotics manufacturers requiring functional safety certification, while sellers are software companies transitioning real-time kernels from the mobile era to wheels and joints. Funding is flowing from automotive pre-authorization to embedded licensing, and then testing in medical and warehousing sectors. Beneficiaries are certified kernels that are difficult to replace with open-source stacks; those under pressure are traders still valuing based on mobile memory. This is a narrative shift, not a return to consumer electronics.
Source: Public Information
ABAB AI Insight
BlackBerry phones died due to app stores and touch screens, while QNX survives in millisecond-level non-crashing control loops. Giamatteo describes cars as "robots with wheels," transferring existing automotive certifications to warehousing, surgical tables, and humanoid frameworks. The backlog of over $900 million indicates that automotive regulatory licenses remain a fundamental base; stating that robots are the fastest-growing segment but rarely discussing it externally suggests a fear of the story outpacing revenue. Caution regarding generative artificial intelligence is to protect the deterministic kernel from being contaminated by probabilistic models.
The capital path involves licensing by vehicle model, designing imports by project, and then using the same certification to sell general embedded systems. The motivation is to turn safety certification into a moat after the mobile brand has zeroed out: it is difficult to obtain the highest-level certifications for automotive and medical at once with open-source and intelligent agents. Collaborating with NVIDIA leverages the wave of computing power without turning BlackBerry into a model company.
This is similar to how Nokia continued to thrive in network equipment after selling off its brand, and also resembles how real-time operating systems like Wind River and Green Hills collect rents in defense and industry. The industry is in a control phase of expanding from automotive to physical artificial intelligence: while models can drift, the execution layer must be certifiable. Whoever first sells "non-crashing" into joint motors will capture the software tax layer that was not obtained during the mobile era.
The essence is the layered replacement after technological substitution. The mobile application layer has been replaced by iOS and Android, but the control layer has not been replaced; it has merely changed carriers. Pricing power has shifted from hardware gross margins to functional safety certificates. The harder the certificate is to replicate, the better positioned the kernel will be in the first round of robotics.
ABAB News · Cognitive Laws
Lost hardware cannot save the surviving kernel
Cars are robots with wheels; certification can be relocated
The layer that cannot crash cannot be replaced by models