The blunt verdict
BlackBerry is a physical-AI company with a certified moat. QNX is the pre-certified safety layer that the world's biggest AI compute company just endorsed on its own keynote stage: on September 22, 2026, NVIDIA's VP of Robotics and Edge AI put up a dedicated "NVIDIA loves QNX" slide at Embedded World North America, presenting QNX as the single pre-certified operating system (IEC 61508 SIL 3, ISO 26262 ASIL D) that unites real-time control with NVIDIA AI compute for humanoids, cobots, and autonomous machines. That is third-party validation of the thesis: as AI moves off the screen and into the physical world, the software underneath has to be certifiable, not just capable, and QNX owns that layer. The September 24 Q2 print is a near-term checkpoint, not the thesis. Wait for the print before sizing any position.
The business engine
BlackBerry is two software businesses sharing one balance sheet. QNX sells the real-time operating system, built to respond instantly and never crash, that sits inside cars, medical devices, and industrial robots. Secure Communications sells ultra-secure messaging and device management to governments and defense agencies. In fiscal Q1 2027 the two engines ran neck and neck: QNX at $72.3M of revenue, Secure Communications at $73.6M, plus $7.0M of licensing. Both core segments grew above 20% year over year, taking total revenue to $152.9M, up 26%. The economics are software economics: QNX posted an 86% adjusted gross margin in Q1 FY27, and each segment converted roughly 27% of revenue into adjusted EBITDA.
The moat: certified, not claimed
QNX's moat is certification, and certification is the hardest moat to fake. A real-time operating system that goes into safety-critical machines must pass years of safety audits before a single royalty is paid. Once a manufacturer certifies QNX into a platform, ripping it out means re-certifying the replacement, which is why the royalty backlog sits near $950M and keeps growing. In April 2026 QNX deepened its NVIDIA collaboration (QNX OS for Safety 8.0 on NVIDIA IGX Thor, and the stock rose about 13% that day), and in September 2026 NVIDIA put QNX on its own keynote slide as the trusted safety layer for physical AI. New optionality comes from SDP 8 and Alloy Kore lifting per-vehicle software content, plus robotics, medical devices, and industrial automation opening royalty streams well beyond autos. The automotive base is the installed foundation; physical AI is the growth frontier.
The ruler: valuation and three-year targets
From the $8.53 reference price, the scenario map is: bear $6.00 (-30%), base $13.50 (+58%), bull $20.50 (+140%). Base-case logic: QNX compounds near 12% annually as physical-AI design wins across robotics, medical, and industrial convert the ~$950M backlog into royalties, alongside continued software-defined-vehicle adoption, Secure Communications grows mid single digits, and incremental royalty revenue flows through at high margins, roughly doubling adjusted EPS from the FY27 guide midpoint of $0.18 to $0.41 by FY29, valued at ~33x, the multiple a net-cash royalty compounder earns, which also marks a long-term technical level the stock has honored before. The bear case is slower royalty conversion and delayed auto recovery, with the turnaround intact and the multiple compressing to 23x. The bull case is QNX becoming the default certified safety layer for physical AI, royalty share climbing toward 72%, and the market awarding 30x on $0.68 of earnings power.
The key compounder mechanism sits underneath those targets: QNX royalties are near 100% margin while development seats and services carry real costs. Royalties are roughly 44% of QNX revenue today; as the ~$950M backlog converts, the royalty share climbs toward 60% and beyond, pushing blended gross margin from ~83% toward ~90%. With largely fixed R&D and negligible capital spending, each incremental royalty dollar falls almost straight to free cash flow, so FCF growth structurally outruns revenue growth. That royalty-driven margin expansion is now built into the 2029 targets rather than left as upside: in the bull case, revenue grows about 57% from FY27 to FY29 while free cash flow roughly triples.
Technical setup: uptrend intact, but extended
The chart is constructive but not a place to chase. The stock is up 103% in twelve months and sits extended above its moving averages, so the trend is intact while the entry is not. Valuation already assumes execution continues, which is why the September 24 print is a useful checkpoint rather than a line on the chart: it confirms whether the growth story is still accelerating.
Risks and kill criteria
Principal risks: royalty revenue follows production volumes, so a broad slowdown in the end markets QNX serves hits directly; government and defense contracts are lumpy, with $220M of ARR as the anchor; the stock's 103% run means the price already knows the good story. A note on auto-cycle timing: our NXP Semiconductors analysis shows the automotive semiconductor recovery slipping a few quarters. This matters less than it used to, because the thesis is now physical AI across robotics, medical, and industrial, not just autos. Near-term QNX revenue is still auto-weighted, so the September 24 print (QNX at or above $75M) remains a checkpoint on whether the royalty line is holding. The kill criteria are simple: a guidance cut would break the CEO's under-promise/over-deliver pattern and force a full re-rating of the thesis.
Download the full report (PDF)
Research and opinion, not investment advice. Do your own due diligence before investing.