Hailo-8
Hailo TechnologiesEdge AI inference accelerator from Hailo Technologies in Israel, delivering up to 26 TOPS at roughly 2.5W typical power. It is sold as a bare chip and in M.2, mPCIe and PCIe card form factors with on-die memory, and supports TensorFlow, TensorFlow Lite, Keras, PyTorch and ONNX through the Hailo Dataflow Compiler and the HailoRT runtime.
hailo.ai answers 403 to every automated request, so the figures were checked against Geniatech's AIM-M-H8 spec table, Waveshare's listing and Hailo's M.2 datasheet mirrored by Premio. The compiler needs registration while the runtime does not, which the openness axis weighs. Verified 2026-08-14 via those listings and the HailoRT README.
Openness
3 high confidence- schematics
- none
- toolchain
- partial (open runtime (HailoRT) + closed/registration Dataflow Compiler)
- datasheets
- brief (public briefs)
- blobs
- required (proprietary silicon + firmware)
- retail
- open_market (global modules)
Proprietary silicon with an open source runtime and public product briefs, but the model-compiler toolchain requires registration and the silicon itself is closed. HailoRT's README carries the open half of the toolchain - "libhailort, pyhailort & hailortcli - distributed under the MIT license", with hailonet under LGPL 2.1 - and the gated half is named by the Model Zoo README, which instructs "Install Hailo Dataflow Compiler ... In case you are not Hailo customer please contact hailo.ai". The same README is why firmware blobs are required: the PCIe driver "includes the Hailo-8 firmware that runs on the Hailo device". The published documentation is Hailo's own Hailo-8 M.2 data sheet Rev 1.1, mirrored publicly by distributor Premio and stamped "Hailo proprietary - unauthorized reproduction prohibited", which is also why no schematics are published at all. Retail availability is Waveshare listing the module at $179.99 with an add-to-cart button. hailo.ai refuses every automated request, so all of this rests on substitute sources.
- https://hailo.ai/products/ai-accelerators/hailo-8-ai-accelerator/ recorded 2026-06-22
official product page: form factors, frameworks, software suite. Not re-fetchable - hailo.ai answers HTTP 403 to every automated request under every header set tried, so the source keeps its recorded read
- https://raw.githubusercontent.com/hailo-ai/hailort/master/README.md recorded 2026-08-14
"HailoRT uses 2 licenses - libhailort, pyhailort & hailortcli - distributed under the MIT license" and hailonet under LGPL 2.1; the PCIe driver "includes the Hailo-8 firmware that runs on the Hailo device"; developer-zone documentation "registration is required"
- https://raw.githubusercontent.com/hailo-ai/hailo_model_zoo/master/README.rst recorded 2026-08-14
"Install Hailo Dataflow Compiler and enter the virtualenv. In case you are not Hailo customer please contact hailo.ai" - the compiler half of the toolchain is customer-gated, while the Model Zoo itself "is released under the MIT license"
- https://premio.blob.core.windows.net/premio/uploads/resource/data-sheet/Hailo-8/Hailo-8%20M.2%20Key%20B_M%20Datasheet%20Rev1.1.pdf recorded 2026-08-14
Hailo's own "Hailo-8 M.2 AI Acceleration Module Data Sheet" Rev 1.1, publicly mirrored by distributor Premio - PCIe Gen 3 2-lane, NGFF M.2 Key B+M, TDP 8.65W, and the footer "Hailo proprietary - unauthorized reproduction prohibited"
- https://www.waveshare.com/hailo-8.htm recorded 2026-08-14
the module on open sale - "Hailo-8 M.2 AI Accelerator Module, based on the 26TOPS Hailo-8 AI processor", SKU 27812, $179.99, Add to Cart
Adoption
4 high confidenceRetail-available modules and a large developer base; the vendor reports 100K+ users and hundreds of deployed customer products, and runs an active GitHub org. That 100K-users figure is not independently derivable - hailo.ai refuses every automated request and no substitute publishes it - so the number is the vendor's own. What the substitutes do establish is the standing the band was placed on: the part is stocked and sold by several independent distributors on three continents (Waveshare at $179.99, Geniatech AIM-M-H8, AAEON/UP, Advantech via Mouser), and HailoRT is a live repo at 215 stars and 80 forks. That is broad standing for an edge accelerator, so level 4 / broad holds. The band rests on that observable distribution rather than on the user count, which is why the unverifiable figure does not move it.
- https://hailo.ai/products/ai-accelerators/hailo-8-ai-accelerator/ recorded 2026-06-22
vendor claim of >100K users and hundreds of customer products. Not re-fetchable - hailo.ai answers HTTP 403 to every automated request - so the claim keeps its recorded read and is not counted as refreshed
- https://www.waveshare.com/hailo-8.htm recorded 2026-08-14
the module in open retail - SKU 27812, $179.99, Add to Cart, listed under Raspberry Pi 5 accessories
- https://www.geniatech.com/product/aim-m-h8/ recorded 2026-08-14
a second, independent module vendor shipping the same silicon - "Geniatech AIM-M-H8 is an M.2 AI acceleration module powered by the Hailo-8 NPU"
- https://api.github.com/repos/hailo-ai/hailort recorded 2026-08-14
HailoRT at 215 stars, 80 forks, last pushed 2026-08-11 - the active developer surface the note describes
Capability
3 high confidenceStrong dedicated CNN inference throughput for the edge, but vision-oriented and limited to models that fit on-die memory. Every element of the recorded value reproduces against substitutes, hailo.ai being unfetchable - Geniatech's AIM-M-H8 spec table reads "AI performance 26 TOPS(INT8)", "AI Model Frameworks Supported TensorFlow,TensorFlow Lite,Keras, PyTorch&ONNX" and "Memory - Processor integration", which is the on-die claim stated as a specification rather than as marketing; Waveshare independently lists 26 TOPS at 2.5W typical. That is one band below the Hailo-10H, which adds a DDR interface and generative-model support.
- https://hailo.ai/products/ai-accelerators/hailo-8-ai-accelerator/ recorded 2026-06-22
26 TOPS, 2.5W typical, on-die memory. Not re-fetchable - hailo.ai answers HTTP 403 to every automated request - so it keeps its recorded read
- https://www.geniatech.com/product/aim-m-h8/ recorded 2026-08-14
spec table - "AI performance 26 TOPS(INT8)", "AI Model Frameworks Supported TensorFlow, TensorFlow Lite,Keras,PyTorch&ONNX", "Memory Processor integration", M.2 Key M, PCIe Gen3
- https://www.waveshare.com/hailo-8.htm recorded 2026-08-14
"AI performance 26 TOPS", "Power consumption 2.5W (Typ.) 8.65W (Max.)", "Interface PCIe Gen3, 4-lane"
Verified 2026-08-14