Ren Malh

[Ren Malh @malby65]

@KanKanandB @sheslee @ashishj150992 @TaiTangOh 💯 and confirmed by the Globe&Mail article today, previous mgmt had the right plan in mind but COVID and OEMs wanting to do everything themselves set them back. Current mgmt is executing perfectly but on very good foundations. That was the vision we, 2020 investors, bought into!

[Ren Malh @malby65]

@KanKanandB 🤘

[Ren Malh @malby65]

@cyberprince_rwo @0xDamien Here https://t.co/L7Pmwo2Q1r

[Prakash @pdamodaran]

The Physical AI Tollbooth: Why BlackBerry QNX Is Positioning for the Comeback of the Decade $BB

Note that i haven't considered Secure communications, Patents, or 5.5 million shares of Arctic wolf . That's all gravy on top.

Most market commentary on artificial intelligence focuses exclusively on digital foundations: large language models, datacenter power grids, and hyper-scale compute clusters.

When attention shifts to physical machines—autonomous robotaxis, factory humanoids, automated mobile robots, and surgical systems , the conversation fixates on hardware platforms like Nvidia or visible fleet operators like Waymo and Zoox.

Nvidia often highlights that its compute stack powers programs across the autonomous spectrum, including Zoox, WeRide, https://t.co/kQZ03cD2UN, Waabi, Mercedes-Benz, Stellantis, and Zeekr.

That raises a core systems question: What is the inescapable common denominator tying these physical machines together?

The answer is not found in neural network training frameworks. It sits lower in the execution stack: BlackBerry QNX.

The Law of Physical AI: The Brain vs. The Spine
Every autonomous vehicle, bipedal humanoid, or heavy industrial robot faces an inescapable architectural split:

The Cognitive Layer (The Brain): Multimodal foundation models, Vision-Language-Action policies, SLAM, and path planning. This layer is data heavy, dynamic, and runs primarily on modified Linux, ROS 2, and runtimes like Nvidia Isaac or PyTorch.
The Deterministic Layer (The Spine): Joint motor inverters, balance loops, torque distribution, drive-by-wire steering, and emergency braking.

Deep learning models are probabilistic. If an inference frame takes an extra 20 milliseconds to process, a video player buffers; but if an operating system suffers that delay during a humanoid stride or vehicle turn, the machine collapses or collides.

Vehicle actuation commands and industrial robotics require hard determinism and functional safety certification: ISO 26262 ASIL-D for passenger vehicles and IEC 61508 SIL 3 for industrial automation.

A monolithic Linux kernel with over 30 million lines of code cannot practically achieve these standards out of the box. Autonomous systems bridge this divide by pairing high performance Linux compute with a pre-certified, deterministic microkernel.

On Nvidia DRIVE platforms (Thor and Orin), QNX OS for Safety provides the pre-certified ASIL-D bedrock required to translate perception into physical vehicle actuation.
In commercial robotaxis using OEM chassis (such as the Zeekr or Jaguar platforms used by Waymo), QNX powers the factory digital clusters, secure gateways, and core domain controllers.
Aside from Tesla, which relies on an in-house, vertically integrated dual-SoC hardware lockstep architecture, the global automotive supply chain runs on standard automotive microkernels.

The Silicon Mandate: Why Global Chipmakers Standardize on QNX 8.0
The semiconductor industry is consolidating tens of legacy electronic control units into centralized, high-performance system-on-chips (SoCs). Silicon vendors cannot sell high end automotive and industrial chips without software that solves multi-core scaling and safety isolation.

Chipmakers worldwide package pre-integrated Board Support Packages and Type-1 Hypervisors for QNX 8.0:

United States: Nvidia (DRIVE Thor/Orin, IGX Thor), Qualcomm (Snapdragon Ride Flex, SA8295P), AMD (Versal AI Edge Gen 2, Kria SOMs), Intel, Texas Instruments.
Europe: NXP Semiconductors (S32G3, i.MX95), STMicroelectronics (Stellar, Telemaco).
Japan & Taiwan: Renesas (R-Car Gen 4), MediaTek (Dimensity Auto).
South Korea: Samsung Electronics (Exynos Auto V920), Telechips (Dolphin 3/5).
China: Horizon Robotics (Journey 6 series), SemiDrive (X9/X10), Black Sesame (Huashan A1000, Wudang C1200), SiEngine (DragonHawk One).

QNX 8.0 was built for this shift, delivering linear CPU scaling up to 64 cores and microsecond interrupt latency. Its Type-1 Hypervisor isolates consumer software, such as Android Automotive OS for in-cabin media, from safety critical tasks running on the same physical die.

Expanding the Footprint: Robotics, Humanoids, and Industrial Automation
Viewing QNX solely through an automotive lens overlooks the broader shift toward physical computing.

In robotics, Nvidia officially integrated QNX OS for Safety 8.0 into its NVIDIA IGX Thor platform, targeting medical devices, factory robotics, and autonomous mobile robots (AMRs). AMD pairs its Versal AI Edge Gen 2 and Kria SOMs with QNX for industrial vision and motion control.

This expansion alters the underlying business model in three key ways:

Higher Selling Prices: Unlike automotive platforms with high volume and tight margins ($5 to $15 per vehicle), industrial machinery, medical robotics, and heavy equipment tolerate software licenses ranging from $50 to $200+ per unit.
Accelerated Development Cycles: Industrial and warehouse automation systems develop on 18 to 24 month cycles, compared to 4 to 7 year automotive timelines, shortening the lag between contract wins and royalty generation.
Single R&D Investment: The identical QNX 8.0 microkernel and hypervisor running on an automotive SoC runs directly on industrial compute modules, enabling expansion across new verticals without duplicate software development.

The Economic Shift: From Single Fees to Multi-Instance Royalties
Historically, embedded software operated on single digit royalties per vehicle (typically $3 to $5 for a basic cluster). The software defined vehicle and centralized physical AI architecture change that profile:

Multi-Instance Deployments: Modern centralized architectures require hypervisor instances, safety OS instances, and telematics layers, pushing automotive royalty capture toward $12 to $20+ per unit.
Compounding Royalty Backlog: BlackBerry's QNX royalty backlog has expanded toward $1 billion, representing contracted design wins already designed into production pipelines.
Operating Leverage: Once software is certified and integrated into silicon reference designs, incremental deployments require minimal capital expenditure. High volume royalty revenue generates gross margins exceeding 80%.

2030 Financial Modeling & Valuation Framework
Assuming BlackBerry common shares trade near ~$7.60 (market capitalization of roughly $4.5 billion on ~587 million basic shares outstanding), we evaluate three potential scenarios through 2030 based on software-defined vehicle adoption and physical AI expansion:

1. Conservative Scenario ($13 to $16 per share)
Assumptions: Automotive QNX monetization remains near $7.50 per vehicle across 60 million annual production cars; robotics expansion remains limited ($120M annual revenue).
Financials: QNX segment reaches ~$680 million in annual revenue; consolidated operating profit nears $380 million; net income approaches ~$310 million (~$0.53 EPS).
Valuation: Applying a 25x to 28x price-to-earnings multiple yields an equity valuation of $7.8 billion to $9.4 billion.

2. Base Scenario ($32 to $38 per share)
Assumptions: Centralized zonal vehicle architectures become standard across major OEMs (ex-Tesla), lifting average automotive royalties to ~$12.00 per vehicle across 70 million cars. Industrial robotics, AMRs, and medical deployments contribute ~$280 million annually.
Financials: QNX segment generates ~$1.30 billion in annual revenue; consolidated operating profit reaches ~$820 million; net income crosses ~$670 million (~$1.15 EPS).
Valuation: Applying a 28x to 32x multiple yields an equity valuation of $18.8 billion to $22.3 billion.

3. Bull Scenario ($65 to $75+ per share)
Assumptions: Multi-instance QNX deployments ($18 to $20+ per car) expand across autonomous and software-defined platforms; factory floor humanoids, heavy autonomous equipment, and high-ASP industrial licenses ($100 to $200+) scale rapidly, generating ~$600 million in robotics revenue.
Financials: QNX segment expands to ~$2.30 billion in annual revenue; consolidated operating profit reaches ~$1.54 billion; net income crosses ~$1.25 billion (~$2.15 EPS).
Valuation: Applying a 30x to 35x multiple yields an equity valuation of $38 billion to $44 billion.

Capital Allocation as an Additional Factor
In scenarios where BlackBerry generates $650 million to over $1 billion in annual free cash flow without requiring heavy capital expenditure for fabs or vehicle assembly, capital allocation strategy becomes an important consideration.

Directing excess cash toward steady share repurchases would reduce the outstanding share count over time, compounding per-share earnings above baseline projections.

The Strategic Takeaway
Technology transitions often reward the infrastructure providers that sit quietly beneath competitive applications.

Autonomous vehicle fleets and robotics platforms will continue to see intense competition, capital burns, and consolidation. However, as long as machines moving through the physical world require certified functional safety, microsecond determinism, and isolated system execution, BlackBerry QNX remains an indispensable component on the global bill of materials.

Standard Regulatory and Financial Disclaimers
This article is published strictly for informational, educational, and analytical purposes and does not constitute financial, investment, legal, or tax advice. It is not an offer, solicitation, or recommendation to buy or sell any security, financial instrument, or commodity, including BlackBerry Limited (NYSE: BB, TSX: BB).

[Ren Malh @malby65]

@sephsephsephs came to realize it's actually not new. The website was launched in January of this year. But never mind, still good find.

[Ren Malh @malby65]

@sephsephsephs Interesting! Very interesting! Thanks for sharing (why would you say it's unofficial - check the copyright at the bottom of the page)

[Ren Malh @malby65]

@trylimitless @trylimitlessfin not a phone company

[Ren Malh @malby65]

@GlenChristy6 EXACTLY ! He's dying to tell us. Any news will be kept for pre-earnings. I am sure one or two AK customers are announced in Sep or Oct. Also, pretty sure $NVDA will take a stake. Just listen to the CEO talking about how not only revenue but investments move share price $BB

[Ren Malh @malby65]

@sadiq_yell @BerryBoysBBHype never happens on fireside chats. Any news will be kept for pre-earnings. Watch this. I am sure one or two customers are announced in September or October. Also, pretty sure $NVDA will take a stake. Just listen to the CEO, he's so antsy about it.

[Ren Malh @malby65]

@sheslee @PassedStranger Based on today's conference I remain pretty sure $NVDA will take a stake in $BB. This and Alloy Kore news will be the catalyst of the remaining quarters this year.

[Ren Malh @malby65]

@pdamodaran @BlackBerry not bad but not super impressed either. When given the chance, I'd like for him to insist on "the modern car is a robot on wheels" narrative (per John Wall)

[Ren Malh @malby65]

@JeysenPlantin @Greenzy15 Do you assume traditional car brands will just disappear? Do you also assume every equipment manufacturer that produces reliable execution in real-time will simply fold to Tesla? Not a chance.

[Ren Malh @malby65]

@KanKanandB brilliant

[Ren Malh @malby65]

@KanKanandB the whole clip is worth a listen but that specific part is very interesting for us $BB

https://t.co/fDKCPxTi16

[Ren Malh @malby65]

@ChapoGrimey lost credibility, unsubscribed

[Ren Malh @malby65]

@IAMessential316 @ChapoGrimey exactly

[Ren Malh @malby65]

@givingemthebiz @billybadass_eth @McBrideLawNYC Sure. While you’re at it, learn about where most of US treasury debt is held. Ignorant fuck.

[Ren Malh @malby65]

@cenkuygur A red card is a red card, accidental or not the Bosnian player could have had his ankle shattered. Like it or not you accept the sentence, grow a pair and move on. You don't get to appeal on it. There are rules and you don't change them in the middle of a world tournament. WTF??

[Ren Malh @malby65]

@cenkuygur Wow wait a minute here Cenk, Belgian here! We weren't mad nor crying about playing the whole US team. No no no no, we knew we'd win - and we did ! We were pissed at the whole Trump Daddy intervention with corrupt to the bones FIFA. BIG DIFFERENCE. Cant believe I am reading this!

[Ren Malh @malby65]

@sheslee @pulkitonline 🫡

[Ren Malh @malby65]

@jarretcoon @Kalshi_Finance You don't get it. The smartphone was not built in a year. Neither was the smart car.