← Research/5/18/2026
Nefarious Trading
Deep Dive · discord.gg/nfrs
Vol. 01 · No. 15
May 02, 2026
BB $5.06 ▲ NVDA $112.34 ▲ QNX 275M+ VEHICLES MBLY $22.40 ▲ GRMN $215.30 ▲ CRWD $385.50 ▲ PLTR $98.20 ▲ NOK $12.28 ▲ BB $5.06 ▲ NVDA $112.34 ▲ QNX 275M+ VEHICLES MBLY $22.40 ▲ GRMN $215.30 ▲ CRWD $385.50 ▲ PLTR $98.20 ▲ NOK $12.28 ▲
$BB · Deep Dive · The Quiet Software Turnaround

BlackBerry isn't a phone company anymore. It runs 275 million cars and just turned profitable.

QNX backlog hit $950M. Revenue up 10%. Eighth straight quarter of margin improvement. The CEO declared the turnaround complete. The Nvidia partnership is real. But the stock already ran 60% in a month and Wall Street consensus sits below the current price.

Price
$5.06
From $3.15 on Apr 6
30-Day Run
+60%
Mid-cycle, not early
Analyst Consensus
$4.75
Below current price
QNX Backlog
$950M
Multi-year contracted
§ 01 — Core Investment Thesis

The turnaround is real. The asymmetry from here is not.

BlackBerry has been a graveyard for value investors for fifteen years. Pivot from phones, pivot to software, miss every quarter, dilute, repeat. April 9, 2026 was the day that ended. Q4 FY26 delivered $156M in revenue, $0.06 EPS, and the eighth consecutive quarter of GAAP profitability improvement. The CEO declared the turnaround complete. The market believed him — the stock ran 60% in three weeks. The thesis is genuinely repaired. The problem now is everyone knows it. Wall Street consensus sits at $4.75 — below the current price of $5.06 — because the easy multiple expansion already happened. This is a real software company with real customers and a real Nvidia partnership. It is also mostly priced.

The TL;DR: Real turnaround, real Nvidia partnership, real $950M backlog. But Street consensus is below current price, growth is 6-11% (not 25%+), and the stock has absorbed most of the rerate. Trade with discipline. The valuation no longer leaves much asymmetric upside.
→ Thesis 01
QNX is a real moat
Safety-certified RTOS in 275M+ vehicles. 68% of EVs. ASIL-D, IEC 62304, DO-178C certifications take 5-10 years to replicate. Not switchable.
→ Thesis 02
The Nvidia deal is structural
Nvidia compute needs a safety-certified OS to ship into regulated industries. QNX is that layer. Already validated in DRIVE AGX Thor and IGX Thor.
→ Thesis 03
Eight straight quarters of profit improvement
GAAP swing from -$79M loss to +$53.2M profit in one year. $432M cash. Buyback authorized. The execution is no longer the question.
§ 02 — What QNX Actually Is

The hidden software layer that 275 million cars run on.

If you do not understand QNX, you cannot value BB. The stock is 90% the QNX thesis and 10% everything else. So here is what it actually is, in language that does not require a CS degree.

QNX is a Real-Time Operating System (RTOS). Not Windows, not iOS, not Linux. A completely different category of OS designed for one specific job: running software where failure can kill people. Cars, medical devices, factory robots, military systems. The boring but critical layer running invisibly inside machines that absolutely cannot crash.

Real-Time vs. Regular Operating Systems

Regular OS
Windows / Mac / Android / Linux
Tasks happen "eventually." Best-effort timing. If your phone freezes for 0.5 seconds, you do not care. Cannot mathematically guarantee that a critical instruction executes within a specific window. That makes it useless for life-or-death applications.
Real-Time OS (QNX)
Deterministic, Provable, Certified
Tasks execute at guaranteed times measured in microseconds. Mathematically bounded worst-case timing. When a crash sensor fires, the airbag deploys in under 30 milliseconds, every time, in every condition. That is what the OS provides.

The Certifications That Are The Moat

QNX is certified to standards that take 5-10 years and millions of dollars to achieve:

ISO 26262 ASIL-D
Automotive functional safety, highest tier. Used in autonomous driving, brake controllers, electric power steering. Replacement cost: 5+ years and major capital.
IEC 62304
Medical device software safety standard. Required for MRI machines, surgical robots, infusion pumps, patient monitoring.
IEC 61508
Industrial functional safety. Used in factory automation, power grids, process control.
DO-178C
Aerospace safety standard. The gold standard for software in aircraft systems.
POSIX-compliant
Standardized developer interface. Means engineers do not need to learn proprietary APIs to build for QNX.

The Toll-Booth Business Model

BB does not sell QNX as a one-time license. They charge a royalty per device that ships with QNX inside it. Every car, every medical device, every robot pays a small fee (typically $5-50 per unit). With 275 million vehicles already running QNX and 68% of new EVs shipping with it, that's a recurring stream that compounds as the installed base grows.

The $950M royalty backlog is contracted future revenue from cars and devices that have not been built yet. As OEMs sign multi-year design wins (Mercedes-Benz, BMW, Volvo, Leapmotor), BB books the future royalty stream as backlog. It is the closest thing to ARR a hardware-adjacent business can produce.

Why The Nvidia Partnership Matters

Nvidia makes the AI compute. The GPU. The brain. But Nvidia chips have no safety certifications. You cannot drop a raw Nvidia GPU into an MRI machine, an autonomous truck, or a surgical robot without an OS layer that handles deterministic timing and isolation. QNX provides that layer.

The integration is now formal: Nvidia DRIVE AGX Thor (autonomous vehicle compute) and Nvidia IGX Thor (industrial/medical edge AI) both ship with QNX OS for Safety 8.0 and the Halos Safety Stack. Validated. Generally available. Shipping product, not press release vapor.

The honest take on QNX: This is genuinely a strong moat business. 45 years old, certified to standards that take a decade to replicate, embedded in 275M+ devices with sticky royalty economics. Hardware-adjacent ARR. The thesis is not made up.
§ 03 — The Nvidia Partnership

This is the real one. Shipping product, not press release.

The BB/Nvidia integration is structurally different from the typical "legacy company gets AI partnership" trade. The product exists, the revenue exists, and the relationship is commercial rather than promotional.

ElementBB / Nvidia (QNX)
PartnerNvidia directly — no shell or rebrand intermediary
Product statusDRIVE AGX Thor + IGX Thor with QNX OS for Safety, generally available
Revenue todayQNX +20% YoY in Q4 FY26, $78.7M record quarter
Customer baseMercedes-Benz, BMW, Volvo, Leapmotor, broader OEM design wins
Equity stakeNone — pure commercial relationship (no dilution risk)
Revenue modelRoyalty per vehicle, sticky for the life of the program

The structural logic: Nvidia builds AI compute, but Nvidia chips ship without safety certifications. To sell into autonomous vehicles, surgical robots, industrial automation, and aerospace, an Nvidia GPU needs an OS layer that handles deterministic timing, isolation, and the certification stack that regulators require. QNX is that layer. BB needs Nvidia compute to handle modern AI workloads on top of safety-certified hardware. Complementary, not competitive.

This is a bet on Nvidia continuing to be Nvidia — the largest AI hardware company in the world, expanding into edge AI in regulated industries — and on QNX continuing to be the default safety RTOS underneath that expansion. As long as both halves of that hold, the partnership compounds.

The honest read: Real product shipping today. Real customers. Real revenue. No equity entanglement. The partnership is the kind of structural integration that's hard to disrupt — not the kind that depends on a single keynote or quarterly press release to keep the narrative alive.
§ 04 — The Numbers

Q4 FY26 was the proof. The full year is more honest.

MetricQ4 FY25Q4 FY26Δ YoY
Total Revenue~$141M$156.0M+10%
QNX Revenue~$66M$78.7M+20% (record)
Secure Communications$67M$72.5M+8%
Adjusted Gross Margin~73%78.2%+500 bps
Adjusted EBITDA$21.1M$36.1M+71%
Adjusted EPS$0.04$0.06+50%
Operating Cash Flow$41.8M$45.6M+9%
QNX Royalty Backlog$865M$950M+10%
Cash + Investments~$280M$432.4M+54%

FY27 Guidance — Real But Modest

MetricFY26 ActualFY27 GuideImplied Growth
Total Revenue$549M$584M-611M6-11%
Non-GAAP EPS~$0.13$0.15-0.19+15-46%
Operating Cash Flow$92M~$100M+9%
Buyback AuthorizationNone6.7M shares~$33M at current price

The Q4 numbers are clean. The full-year numbers are more sober: FY26 total revenue grew only 3% ($533M to $549M). FY27 guidance of 6-11% growth is real, but it is not the 25%+ that justifies a software multiple expansion.

The growth engine is QNX (+14% for full FY26, +20% in Q4). Secure Communications returned to growth (+8%). UEM (the legacy device management line) declined. The "growth company" narrative is true for two of three segments. The third is still bleeding.

What "Rule of 40" means here: Management celebrated QNX hitting Rule of 40 — the SaaS metric where revenue growth + EBITDA margin should exceed 40%. QNX did it (20% growth + 27% EBITDA margin = 47). The full company did NOT — 10% revenue growth + 23% EBITDA margin = 33. Real software companies generating premium multiples typically clear 50-60. BB is on the path. It is not there yet.
Join the Community

Trade ideas like this, before they hit the timeline.

Join Discord
discord.gg/nfrs · @johnnylixf
⬢ The Community
24/7
Active Discussion Across Time Zones
100%
Free · No Paid Signals
3+
Deep Dives Per Month
Real Trade Reviews & Postmortems
"Trading is a solo sport. The community part is what makes you better at it."
§ 05 — The Wall Street Reality

Analyst targets are below the current price.

This is the part of the story the bull narrative skips past. After the Q4 beat, after the Nvidia announcement, after the 60% run, here is what professional analysts who model this company for a living are saying:

Analyst Targets — Post-Earnings (April 2026)

CIBC $6.00 Highest target. Buy. Todd Coupland.
RBC Capital $4.50 Hold. Paul Treiber. -11% from current.
Canaccord $4.40 Hold. Lowered post-earnings. -13% from current.
TD Cowen ~$4.50 Hold. Cited "reinvestment risks."
Median Consensus $4.75 10 analysts. 1 Buy, 7 Holds, 0 Sells. Below current.

The pattern matters. After the Q4 earnings beat, Canaccord LOWERED their target from $4.60 to $4.40. RBC kept Hold. TD Cowen maintained Hold and explicitly cited "reinvestment risks." Only CIBC sees meaningful upside, and even their $6 target is just 19% above current price.

The retail bull case sits roughly 4x above the highest professional target. That is not a small gap. The professional analyst community watched the same earnings, modeled the same QNX growth, and concluded $4.40-$6.00 is fair value. The gap between consensus and the more aggressive retail price targets is the asymmetric risk.

The honest reading: When the Street is below the current price after a beat, momentum is doing the heavy lifting, not fundamentals. That can extend further but it usually does not extend much further. Position accordingly.
§ 06 — Competitive Position

The safety RTOS market is a three-name oligopoly. QNX is the leader.

Safety-certified RTOS is not a crowded market. Building one takes a decade and tens of millions of dollars in certification work. Once an OEM picks an RTOS for a vehicle program, the switching cost is so high it almost never happens. The result: three Western pure-plays and a handful of Linux/Android contenders trying to break in. Here is the actual landscape.

CompanyOwnershipAuto OS ShareSafety CertsPosition
QNX (BlackBerry)Public (NYSE: BB)~37% (2025)ASIL-D, IEC 62304, DO-178C, IEC 61508Leader. 275M+ vehicles. 68% of EVs.
Wind River (VxWorks)Aptiv (NYSE: APTV) — bought 2023 for $3.5BTop-3ASIL-D, DO-178C, IEC 61508Strong in aerospace/defense. Volvo ADAS win 2025.
Green Hills (INTEGRITY)PrivateNicheASIL-D, DO-178C, EAL 6+Defense/avionics-heavy. Smaller auto footprint.
Android AutomotiveAlphabet (GOOG)Rising in IVINone for safety layerInfotainment only. Runs on top of QNX/Linux for ADAS.
Embedded Linux (AGL etc)Open sourceGrowingLimited (ELISA in progress)Cost-driven. Not certified for ASIL-D.

The Honest Wind River Comparison

Wind River is the closest direct competitor. Aptiv paid $3.5B for it in 2023 — at roughly $400M of revenue, that's an ~8.75x sales multiple paid by a strategic acquirer. BB's entire enterprise value at $5.06 is ~$3B for ~$200M of QNX revenue (so ~15x QNX sales). On that comparison, BB is more expensive — but BB is also growing QNX faster than Wind River grew, and BB still has Secure Comms cash flow on top.

Wind River won a major Volvo ADAS deal in June 2025 — that matters. It's the first big-ticket auto win for VxWorks against QNX in a while, and it tells you the duopoly is genuinely competitive at the OEM level. QNX is still the share leader, but it is not unchallenged. Green Hills sits in the corner with the highest-assurance avionics business but a small auto presence. Android Automotive is a marketing distraction — it runs on top of a safety RTOS, not in place of one.

The clearing argument: Three certified safety RTOS players compete for OEM design wins. QNX leads in autos (~37% share), Wind River leads in aerospace, Green Hills leads in defense. The acquirer math from Aptiv's Wind River deal is the cleanest comp BB has — and it implies a strategic floor under QNX value if BB ever sold the unit.
§ 07 — Scorecard

Real turnaround. Mostly priced asymmetry.

Bull Case

  • Q4 FY26 was a clean beat. $156M revenue (+10%), $0.06 EPS beat consensus, 8th straight quarter of profitability improvement.
  • QNX moat is real. 275M+ vehicles, 68% of EVs, certifications that take 5-10 years to replicate. Not switchable.
  • Nvidia partnership is structural and shipping. DRIVE AGX Thor and IGX Thor with QNX OS for Safety, generally available.
  • $950M royalty backlog. Multi-year contracted revenue visibility. Closest thing to ARR a hardware business can produce.
  • $432M cash, net cash positive. No dilution risk. Buyback authorized.
  • Robotics/medical optionality. Physical AI wave applies QNX certifications to a TAM 5-10x larger than auto alone.
  • Secure Communications back to growth. Digital sovereignty + defense budgets are tailwinds.
  • Buyout optionality. Strategic value to Nvidia, Marvell, or a major OEM. Not happening tomorrow but possible.

Bear Case

  • Stock ran 60% in 30 days. Mid-cycle, not early. Most easy multiple expansion already priced.
  • Wall Street consensus $4.75 — BELOW current price. Canaccord LOWERED target post-earnings. RBC, TD Cowen on Hold.
  • Full-year revenue grew only 3%. Q4 +10% was the standout. Annual growth is well below software peer averages.
  • FY27 guide is 6-11%. Not 25%+ growth. Hard to justify 54x P/E without faster growth.
  • UEM/Licensing declining. Legacy line still bleeding while QNX grows. Drag on consolidated metrics.
  • P/E 54x on slow-growth software. CRWD trades at similar multiple but grows 25%+. BB at 6-11% growth is rich.
  • "Royalty backlog" is misunderstood. $950M is multi-year (5-7 years). Annual conversion is far smaller. Bulls treat it as imminent.
  • Retail price targets ($20+) sit 4x above the highest analyst. Crowded long with no professional support at those levels.
§ 08 — Price Targets

Where it can go. Where it probably stops.

Honest scenario map. The bull case is real but bounded. A $20 print doesn't math at this revenue trajectory.

Bear · 3-6mo
$3.75
−26%
Q1 FY27 in-line rather than beat. Auto seasonality. Multiple compresses to 40x as growth recognition slows. Most likely bear scenario.
Base · 6-12mo
$5.50
+9%
Slightly above consensus. Q1-Q3 maintain mid-teens QNX growth. Secure Comms holds 8%. Most likely landing zone if execution holds.
Bull · 12-18mo
$7.00
+38%
CIBC target plus modest premium. QNX hits 20%+ for full FY27. Major analyst upgrades to Buy. Robotics revenue line emerges.
Stretched · 24mo
$10
+98%
Multi-year rerate to 8x sales as real growth software. Buyout speculation. Robotics becomes meaningful revenue. Possible but requires sustained execution.
§ 09 — My Take

The thesis is real. The valuation has caught up.

BB is a real turnaround with a real Nvidia partnership and a real QNX moat. Q4 FY26 was the proof print: $156M revenue, $0.06 EPS, eight straight quarters of profitability improvement, and a CEO who can finally say the turnaround is complete with a straight face. The market believed him — the stock ran 60% in three weeks.

The defensible read: QNX is genuinely the safety layer underneath Nvidia's edge AI strategy. The royalty model produces sticky, compounding, hardware-adjacent ARR. The certifications create a 5-10 year switching cost moat. The auto + medical + industrial + robotics TAM is real. The CEO is correct that the turnaround is complete.

The honest read: at $5.06, the stock is pricing all of that. Wall Street consensus is $4.75. Growth is 6-11%, not 25%. UEM is still declining. The aggressive retail price targets are framing more than analysis — the math doesn't support a near-term $20 print at this revenue trajectory.

This is a trade with discipline. Not a position you sleep on at $5+. Not a $20 multi-bagger.

Long BB on pullbacks to $4.50-4.80. Not chasing here.

Entry zone: $4.50-4.80 (analyst consensus zone, where Wall Street and the chart agree). Trim 50% at $6.00 (CIBC target). Trim another 25% at $7.00. Hard stop on any break of $4.20. Adding only on a pullback with volume contraction, not on continuation. Position size MEDIUM. Not a small-cap meme but not a high-conviction multi-bagger either. The Q1 FY27 print (June 2026) is the next real catalyst — either confirms the trajectory or breaks it.

"This is the kind of trade where the analyst targets matter more than the narrative."
One Last Thing

If this report saved you from chasing the top — come trade with us.

Join Nefarious Trading
discord.gg/nfrs · @johnnylixf
⬢ What You Get
100%
Free · No Paid Signals · No Upsells
3+
Deep Dives Per Month Like This One
24/7
Live Trade Discussion & Setups
Postmortems · No Cherry-Picked Wins
"Read the analyst targets. Then come check what the room thinks."