All Projects
Pitch DeckSource
ESA EXPRO+ Proposed

P.R.I.S.M.

AI that discovers materials faster than humans can test them

Autonomous AI-driven discovery engine for accelerated materials synthesis, targeting refractory high-entropy alloys for space propulsion

July 20251 contributor
AIMaterials DiscoveryRHEAESA EXPRO+Additive Manufacturing
P.R.I.S.M. AI discovery pipeline
10⁶+
Candidates screened
4
Autonomous subsystems
<1 hr
Per screening cycle
ESA EXPRO+
Proposal submitted

P.R.I.S.M.: Autonomous Materials Discovery

PRISM (Platform for Research in Intelligent Synthesis of Materials) is an autonomous discovery engine that replaces the traditional multi-year materials R&D cycle with AI-driven computational screening and robotic synthesis. The initial target: refractory high-entropy alloys (RHEAs) for oxygen-rich preburners in liquid rocket engines, replacing legacy materials like Monel K500.

How It Works

The system runs a dual-loop workflow — a computational inner loop screening thousands of candidates per hour, and a physical outer loop validating top picks through robotic synthesis and hot-gas testing.

The Evolver (ACE Brain)

The orchestrating reasoning model. Uses Agentic Context Engineering to build and refine a "playbook" of discovery strategies, balancing exploration and exploitation through a Generator, Reflector, and Curator — without suffering context collapse.

The Mutator Fleet

Specialized AI sub-agents that propose candidate modifications:

  • Composition Mutator — elemental substitutions
  • Process Mutator — synthesis protocol optimization
  • Thermo Mutator — thermodynamic querying and phase stability checks

The Evaluator

Multi-tiered screening pipeline providing the fitness score. Fast surrogate models (MACE-MH-1 equivariant GNN) and ab initio DFT/CALPHAD checks triage candidates computationally, then route the best to physical A-Lab robotic synthesis for ground-truth data.

Materials Knowledge Graph

The system's active memory. Ingests heterogeneous data from scientific literature, patents, and experiments. The Evolver queries it to break out of search stagnation and formulate new hypotheses.

Consortium

Structured as an ESA EXPRO+ proposal with a 12-month de-risking timeline:

  • Bimo Tech (Prime) — AI architecture, PRISM digital design, surrogate model development, Laser Powder Bed Fusion (LPBF) synthesis of down-selected alloys
  • Fraunhofer IAPT — AI-based Design for Manufacturability (DfM) scoring, R&D LPBF process calibration, defect evaluation
  • ArianeGroup — Industrial end-user requirements, physical validation via ERBURIGK hot-gas oxidising facility, benchmarking against Monel K500
  • amsight — Data integration, structuring manufacturing and qualification data into machine-readable formats for model retraining

Status

  • PRISM-alpha open-sourced July 2025
  • ESA OSIP market research study positively evaluated
  • ESA EXPRO+ proposal (Ref: ESA AO/1-13197/25/FR/LCF) submitted under FIRST! Simulation & Intelligence Technologies
  • Active data pipelines with NCBJ and IPPT PAN (Poland) for scaling and subscale testing
  • Pending Phase 1 kick-off: Cold Start bootstrap using transfer learning from near-zero domain data