Company profile

Alp Lab Edge-AI

Open-standard Edge AI modules and unified SDK for embedded systems.

alplab.aiProfile compiled July 202612 source pages read
Category
Chips & hardware
Headquarters
Västerås, Sweden
Sells to
Developers
Business model
Hardware, Services & consulting
Deployment
Edge
Pricing
One-time purchase for dev kits and modules; fixed-quote NRE for engineering services.
Builds own models
No — builds on existing models
Modalities
Sensor, Video

Alp Lab specializes in embedded software and hardware innovation, offering open-standard Edge AI modules (E1M™) and a unified SDK (Alp SDK™). Headquartered in Sweden, the company provides vendor-neutral, pin-compatible modules from various silicon vendors like Alif, NXP, Renesas, and DeepX, designed for high performance, flexibility, and open-source principles in Edge-AI and IoT markets. Their solutions aim to simplify development by providing a single SDK and API across different hardware, ensuring industrial reliability and long-term availability for harsh environments. Alp Lab also offers engineering services including carrier-board design, BSP customization, and AI model bring-up.

  • E1M™ (Edge-1 AI Module)A platform of open edge AI, vendor-neutral, pin-compatible modules from Alif, NXP, Renesas, DeepX, available in two footprints with one contract and one SDK. It offers AI compute ranging from 0.25-0.5 TOPS for always-on sensing to 33 TOPS for heterogeneous compute, all using the same unified Alp SDK and C/C++ API.
  • Alp SDK™A unified C/C++ API and SDK that wraps each vendor SDK on top of chosen OS (Zephyr, Yocto, bare-metal), allowing a single application to build for any E1M module. It includes AI Framework for on-device inference, Ethos-U, DRP-AI3, TFLM runtime, and Dev Tooling for build, orchestration, and flash.
  • E1M-AENAn E1M module for always-on AI at sub-1 mW, featuring MCU+MPU (2xM55 + opt. A32) and NPU (2x Ethos-U55) with 0.25-0.5 TOPS AI performance. It comes in a 35x35 mm form factor and is ideal for wake-word, vibration anomaly, and ultra-low-power sensing.
  • E1M-N93An E1M module for industrial real-time AI, featuring 0.5 TOPS AI, 2xA55 CPU, Ethos-U65 NPU, and TSN (2x 1 GbE TSN). It comes in a 35x35 mm form factor.
  • E1M-X V2NAn E1M-X module with 4 TOPS AI, 4x A55 CPU, and DRP-AI3 NPU, in a 45x65 mm form factor. It includes MIPI, PCIe, and USB3 interfaces.
  • E1M-X V2HAn E1M-X module with 80 TOPS (sparse) AI, 4x A55 + 2x R8 CPU, and DRP-AI3 NPU, in a 45x65 mm form factor. It includes MIPI, PCIe, USB3, and 2x ETH interfaces.
  • E1M-X V2N+M1An E1M-X module combining RZ/V2N with DeepX DX-M1 for 29 TOPS combined AI in a 45x65 mm form factor.
  • E1M-X V2H+M1An E1M-X module combining V2H with DeepX DX-M1 for 33 TOPS dense (105 TOPS sparse) combined AI in a 45x65 mm form factor.
  • Smart Eye CameraAn industrial Edge-AI Camera powered by the E1M-X V2N+M1 SoM. It's a lightweight, all-in-one smart vision device for high-performance, low-power AI processing in manufacturing, robotics, warehouse automation, and workplace safety, housed in a compact, rugged IP67 housing. Configurable from 4 to 33 TOPS by swapping the underlying SoM.
  • Carrier-board design serviceFull carrier board design from sketch to first-spin gerbers, including schematic capture, PCB layout, SI simulation, and first-article bring-up with firmware on the Alp SDK.
  • BSP customization serviceTailored Yocto, Zephyr, or bare-metal BSP for specific peripheral sets and production lines, including boot-time optimization, production-flash provisioning, and EEPROM identity.
  • AI model bring-up serviceCompiling trained models onto SoM accelerators (Ethos-U, DRP-AI3, DEEPX DX-M1), quantization, INT8 calibration, real-world latency and power profiling, and integration through the Alp SDK inference dispatcher.
  • Open edge AI, vendor-neutral
  • Pin-compatible modules from Alif, NXP, Renesas, DeepX
  • Two footprints, one contract
  • One unified SDK (Alp SDK)
  • Hardware-agnostic
  • AI-optimised
  • Industrial-grade reliability (-40 °C to +85 °C operating temperature, vibration resistance, 10-year minimum supply)
  • Scalable AI performance (0.25-33 TOPS)
  • Unified C/C++ API across all modules
  • Support for Zephyr, Yocto, bare-metal OS
  • Flexible Edge AI system-on-module architecture
  • Full footprint compatibility for seamless upgrades/downgrades
  • Open standard under CC BY-SA 4.0
  • Deterministic dual-OS architecture (Yocto + Zephyr) on a single module
  • Secure boot + DICE attestation supported on all modules
  • Compliance with RoHS 3, REACH SVHC, POPs, California Prop 65, Halogen-free, Conflict Minerals (Dodd-Frank §1502), WEEE 2012/19/EU
  • Autonomous flight for drones
  • Object detection for drones
  • Mission planning for drones
  • ROS 2 perception on NPU for robotics
  • Real-time motion control on R-cores for robotics
  • Visual inspection in industrial automation
  • Predictive maintenance in industrial automation
  • Process optimization at the edge in industrial automation
  • AI-powered crop monitoring in agriculture
  • Autonomous farm equipment
  • Smart grid monitoring in energy
  • Solar tracking in energy
  • Equipment diagnostics at the asset in energy
  • People counting in smart infrastructure
  • Traffic flow monitoring in smart infrastructure
  • Energy monitoring in smart infrastructure
  • Wake-word detection
  • Vibration anomaly detection
  • Ultra-low-power sensing
  • High-throughput inference at the edge
  • Workplace safety
  • Collision avoidance for robotics
  • Manipulation for robotics

Alp Lab provides open-standard Edge AI modules (E1M™) and a unified SDK (Alp SDK™) for embedded software and hardware innovation. Their modules are hardware-agnostic, AI-optimized, and industrial-grade, designed to run complex AI models locally with minimal latency. The Alp SDK allows developers to write firmware once and deploy it across various E1M modules, wrapping vendor SDKs on top of chosen OS (Zephyr, Yocto, bare-metal). They support on-device inference with various AI frameworks and accelerators.

Tech named: E1M™ (Edge-1 AI Module), Alp SDK™, Alif Ensemble E3–E8, NXP i.MX 93, Renesas RZ/V2N, Renesas RZ/V2H, DeepX DX-M1, Ethos-U55, Ethos-U65, DRP-AI3, TFLM runtime, C/C++ API, Zephyr, Yocto, bare-metal, Arm CMSIS, Cortex-M55, Cortex-A32, Cortex-A55, Cortex-M33, Cortex-R8, YOLOv8-tiny, 1D-CNN, ROS 2

  • IT Services and IT Consulting
  • Edge-AI
  • IoT
  • Robotics
  • Smart Manufacturing
  • Drones & UAVs
  • Industrial Automation
  • Agriculture
  • Energy
  • Smart Infrastructure
  • Warehouse Automation
  • Open-standard and vendor-neutral approach, allowing silicon changes without base board redesign
  • Unified SDK and C/C++ API across diverse silicon vendors, streamlining development
  • Industrial-grade reliability and long-term availability (10-year minimum supply) for harsh environments
  • Pin-compatible modules enabling seamless upgrades or downgrades without hardware redesign
  • Focus on customer success by providing free schematic reviews and flexible engineering services with fixed-quotes and no royalties
  • Deterministic dual-OS architecture on a single module for complex applications like robotics
  • Comprehensive compliance and regulatory support for module-level declarations

This profile was compiled from Alp Lab Edge-AI's own public pages in July 2026 and reflects what the company states about itself — not an endorsement or an independent audit of those claims. Facts are extracted with AI and filtered by an automated check that drops any named product, customer or certification missing from the source pages. Full method. Something out of date? Tell us.