NEOMA launches hosted MCP server for P-CAM analog silicon Ising solver
NEOMA released v5.4.0 of its P-CAM hardware platform, now reachable via a hosted MCP endpoint that lets AI agents like Claude Desktop, Cursor, or n8n call physical silicon directly over HTTP. The release adds tools for nonlinear equilibrium solving across seven domains, a 64-spin analog Ising Hamiltonian optimizer, and a cryptographic Silicon PUF verification tool, alongside a new Python SDK (pcam-sampler) on PyPI with a REST endpoint for Ising solving.
GoKawiil's interpretation of the reporting above, not reported fact.
Exposing specialized analog hardware through the Model Context Protocol could let AI agents offload physics-based optimization tasks—like thermal modeling or combinatorial scheduling—to dedicated silicon rather than simulating them in software, which the company frames as faster and free of pseudo-random number generator artifacts. The hardware provenance verification tool suggests an interest in letting users confirm results genuinely came from physical hardware rather than a simulated stand-in, though claims about speed and determinism rest on the vendor's own benchmarks.
- NEOMA now offers a zero-install hosted MCP server connecting AI agents to its P-CAM analog silicon hardware
- The release adds a 64-spin Ising Hamiltonian solver, multi-domain equilibrium solver, and a cryptographic hardware-verification tool
- A new Python SDK (pcam-sampler) and REST API now support direct Ising optimization calls outside the MCP interface
Source: github.com, 2026-11-16
Published there as: “Show HN: Neoma P-CAM – Sub-millisecond physical equilibrium and 64-spin Ising”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.