Skip to content

§ Why open

Open beats locked-in for edge AI.

Most edge AI modules tie your product to one silicon vendor — their pinout, their SDK, their roadmap, their supply. E1M is an open standard built to remove that risk.

The lock-in problem

Pick a typical edge AI System-on-Module and you inherit three dependencies at once: a proprietary pinout that fixes your carrier-board design to one family, a vendor SDK that fixes your firmware to one toolchain, and a supply roadmap you don't control. When the vendor raises prices, hits an allocation shortage, or ends a part, your only options are a board respin or a firmware rewrite — usually both.

What an open standard changes

E1M fixes the contract between the module and your carrier board — the pinout, the mechanical envelope, and the electrical interface — and publishes it under CC BY-SA 4.0 onGitHub. Two footprints cover the range: 35 × 35 mm (312 pads) and 45 × 65 mm (496 pads). Any module that conforms drops onto the same base board.

Swap silicon in one line

Because the SDK abstracts the vendor layer, changing the silicon under your product is a configuration edit, not an engineering project. Set the module SKU in board.yaml and the SDK rebuilds your application — Yocto on the application cores, Zephyr on the real-time companion, IPC stitched in — against the same<alp/*.h> API.

Where to go next

Read theE1M open standard, browse theE1M modulesthat implement it, or see how E1M compares toother edge AI platforms.

Frequently asked questions

What does vendor-neutral edge AI mean?
Vendor-neutral means your product is not tied to one silicon supplier. With the open E1M standard the carrier-board pinout, mechanical envelope, and electrical interface are fixed and published, and the Alp SDK exposes one C/C++ API across Alif, Renesas, NXP, and DeepX silicon — so you can change vendor without redesigning your board or rewriting your firmware.
How is E1M an open standard?
The E1M specification — pinout, two mechanical footprints, and electrical interface — is published publicly under CC BY-SA 4.0 on GitHub. Anyone can build a conforming module or carrier board. It is a standard, not a single vendor’s product line.
What is a second source for a System-on-Module?
A second source is an alternative supplier you can switch to without redesigning. Traditional SoM "pin-compatibility" only spans one vendor’s own family on the same silicon. E1M is pin-compatible across silicon vendors, so a different module — even on different silicon — drops onto the same carrier board.
How do I avoid edge AI vendor lock-in?
Choose an open hardware standard plus a portable software layer. On E1M, changing the silicon under your product is a one-line som.sku edit in board.yaml — the SDK rebuilds for the new vendor against the same <alp/*.h> API, instead of a board respin and a firmware rewrite.

Build on the open E1M standard.