No cloud in the loop
Inference completes on the device. There is no upload path to secure, no telemetry to disclose, and no service to keep paying for.
Solutions
Fixed power budget. Unreliable or absent network. A decision that has to land before the next control cycle. Wherever those three meet, moving weights across a bus stops being affordable.
Fixed power, no network, hard deadline
The same three constraints turn up in every industry on this page, which is why one architecture answers all six.
The decision stays on the machine
Perception and response both complete on the device, so a product works in a tunnel, at sea, and on a line that will not touch the public internet.
License the engine, or take the part
Into your own silicon as IP, or as a module with the toolchain around it. Both paths start with the same workload review.
Inference completes on the device. There is no upload path to secure, no telemetry to disclose, and no service to keep paying for.
Weights are non-volatile and live in the array. A node can sleep for days and be ready on the first sensor event.
Work is distributed across many small arrays. A damaged region costs precision in one place instead of taking the system down.
The engine is a co-processor on your existing MCU or SoC. It takes the perception workload; your host keeps the application.
Automotive
Sensor fusion, driver monitoring and predictive diagnostics that run inside the ECU on a power budget a vehicle can actually spare, with a wake-up that costs nothing, because the weights never left the array.
Not on the list
If your workload is dominated by matrix-vector products, runs continuously on a constrained supply, and cannot depend on a network round-trip, it belongs in this set. Whatever the industry is called. An evaluation will tell you within weeks, including when the answer is no.
Evaluations start with your workload, your power envelope and your deadline, not with a datasheet.