Zephyr RTOS for Automotive
A decade of open-source maturity. 1,500+ contributors. 600+ supported boards. Zephyr is one of the most actively developed RTOS projects - and Linumiz is bringing it to automotive.
10 Years of Zephyr
Since its founding as a Linux Foundation project in 2016, Zephyr has grown from a lightweight IoT kernel into a full-featured RTOS powering devices across industrial, medical, consumer, and now automotive domains.
Project Founded
- Linux Foundation Collaborative Project
- Wind River donates the Rocket kernel
- First target: x86, ARM Cortex-M
Connectivity
- Thread, Bluetooth Mesh, USB stabilized
- 200+ board targets
- First commercial deployments
LTS & Maturity
- Long-Term Support program launched
- Production-grade kernel stability
- 400+ board targets
Industry Adoption
- 600+ board targets
- Platinum members: Intel, Nordic, NXP, TI
- 1,500+ contributors
CAN & Automotive Interest
- CAN / CAN FD support matures in Zephyr
- TriCore architecture interest in the community
- ISO 26262 safety discussions start
Automotive Stack
- Linumiz: production CAN FD, SPI, I2C, MPU, WDT
- Labwire: CI/CD + remote lab on real hardware
- auto.linumiz.com launched
- Upstream PR flow to Zephyr project
Why Zephyr for Automotive
- Open source, vendor neutral. No license fees, no vendor lock-in. OEMs and Tier-1s can inspect, modify, and certify the kernel themselves.
- Deterministic, real-time. Tickless kernel with configurable scheduling policies. Bounded interrupt latency. Designed for hard real-time from the start.
- Hardware abstraction done right. Device Tree-driven configuration. Peripheral drivers follow a consistent API (CAN, SPI, I2C, GPIO, PWM, ADC). Swap silicon without rewriting application code.
- Safety-ready architecture. MPU-based memory isolation, user/kernel separation, and deterministic trap handling. The building blocks for functional safety qualification.
- Massive ecosystem. 1,500+ contributors, 80+ silicon vendors, mature CI (Twister), and an active community. Fixes and features land upstream fast.
Zephyr vs. AUTOSAR
Zephyr and AUTOSAR serve different layers of the automotive stack. They're complementary, not competing.
| Zephyr RTOS | AUTOSAR Classic | |
|---|---|---|
| License | Apache 2.0 (free) | Consortium membership |
| Scope | RTOS kernel + drivers | Full middleware stack |
| Footprint | ~8 KB min | 100+ KB typical |
| Vendor lock-in | None | Varies by stack vendor |
| Safety cert | IEC 61508 SIL 3 pathway | ASIL-D (mature) |
| Best for | Sensor ECUs, actuators, zone controllers | Complex ECUs, gateway, powertrain |
For many automotive use cases - sensor fusion, body control, CAN gateway - Zephyr delivers the right abstraction at a fraction of the cost and complexity.
Industry Adoption
Independent Linux Foundation survey of 413 organizations - Zephyr Turns 10, March 2026.
commercial products (US/CA)
only 1% expect to decrease
for 5 to 10+ years
ecosystem maturity (46%)
Biggest reported impact of adopting Zephyr
Source: LF Zephyr Turns 10 (Fig 19), March 2026.
Zephyr Speaks Automotive
Determinism, isolation, and a safety pathway are baseline kernel properties - not add-ons.
Deterministic timing
Tickless kernel, bounded interrupt latency, Zero-Latency Interrupts (ZLI).
Spatial isolation (FFI)
MMU / MPU memory domains for Freedom From Interference - the core ISO 26262 requirement.
Predictable scheduling
Cooperative + preemptive. Configurable policies for safety-critical task priority.
Functional safety pathway
SWG targets IEC 61508 SIL 3 via Route 3S. Maps to ISO 26262 (road) and ISO 25119 (agricultural).
Disciplined C
MISRA C:2012 guidelines applied to safety-critical subsystems.
Independently benchmarked
thread_metric ships in-tree. Zephyr 4.1 on par with ThreadX, ahead of FreeRTOS.
Silicon Platforms
Linumiz delivers Zephyr enablement across multiple silicon families - from safety-critical powertrain controllers to cost-optimized body electronics.
Infineon AURIX (TriCore)
32-bit multi-core MCU for powertrain, chassis, and safety-critical ECUs. Hardware lockstep, MPU isolation, CAN FD up to 5 Mbps.
- TC3x and TC4x architecture ports
- CAN, SPI, I2C, UART, GPIO, WDT, MPU, clock, timer
- Upstream to mainline Zephyr in progress
- Safety artifacts developed to ISO 26262 ASIL D process
Infineon Traveo T2G (Arm)
Arm Cortex-M7 + M0+ multi-core automotive MCU for instrument clusters, body control, and gateway applications.
- CYT4DN (Cluster 2D) and CYT4BFC (Body-High)
- ADC, CAN, DMA, Flash, I2C, I2S, MBOX, PWM, SPI, WDT
- Multi-core boot with M0+ clock ownership
- Upstream submission in review
TI MSPM0 (Arm)
Arm Cortex-M0+ for cost-optimized body electronics - window lifters, seat adjusters, lighting, TPMS. AEC-Q100 Grade 1, ASIL B.
- MSPM0G and MSPM0L series
- CAN, SPI, ADC, DAC, PWM, DMA, GPIO, RTC, AES, WDT
- Linumiz engineer is the upstream Zephyr maintainer for MSPM0
- Multiple drivers merged into mainline Zephyr
What Linumiz Delivers
We don't just port an RTOS - we deliver a complete automotive development platform, from kernel bring-up to safety certification:
Silicon Bring-up
Architecture ports, peripheral drivers, board support packages. Three silicon families actively supported across Zephyr and NuttX.
Safety Engineering
MPU isolation, trap handling, watchdog integration. SEooC safety artifacts for ISO 26262 and IEC 61508 qualification.
Upstream Maintenance
Linumiz engineers maintain silicon platforms in mainline Zephyr. Contributions across Zephyr, Linux kernel, U-Boot, and NuttX.
CI/CD + Remote Lab
Hardware-in-the-loop testing on real boards. Automated flash-and-verify. Remote board access from anywhere via Labwire.
Beyond Zephyr
The same silicon expertise and tooling methodology extends to other RTOS ecosystems where customers need it.
Apache NuttX on AURIX TriCore
Full NuttX RTOS port for Infineon AURIX - architecture support, peripheral drivers, and board enablement for both TC3x and TC4x families. Monthly releases since November 2025. Targeting the Asian automotive market where NuttX adoption is growing rapidly.
Upstream to Apache NuttX planned. Same engineering team, same silicon, different RTOS.