Seed Materials Informatics Project

Overview

Seed Materials Informatics (SMI) is one of the core technologies Space Seed Holdings develops in-house. Drawing on the support of generative AI and other tools, we take the many development results reported and accumulated over time and organise them into “seed materials” — the seeds of manufacturing technology — bringing them into a state that is easy to verify and to propose. SMI is the informatics platform built for this purpose.

Left as they are, development results for both materials and biomolecules remain scattered data. SMI aims to be a common mechanism that organises them across domains and shapes them into a form ready for the next round of verification or proposal. We position it as a second core technology, alongside Space Agent (space × autonomous experimentation).

What we are working towards

SMI is being developed along two axes of differing character.

• Materials (sintering) axis — we apply materials informatics (MI) to “recipes for sintering,” beginning with alloys, to organise the relevant information. The aim is to narrow promising candidates in a data-driven way and then shape them into a form that is easy to carry forward into downstream sintering verification.
• Bio axis — we are building a mechanism that supports, from the information side, the development of new modalities (means of treatment or prevention) that address human health challenges. Across steps such as drug repositioning, formulation and molecular design, we aim to assemble candidate approaches and bring them into a form that can be proposed.

What the two axes share is the idea of organising development results as information so that they are easier to verify and to propose. Whether the subject is a material or a biomolecule, SMI’s core lies in building a cross-cutting base that narrows candidates in a data-driven way and connects quickly to the next move.

Technical grounding

The basic idea of materials informatics is to learn regularities from past development data and to narrow down promising options from a vast candidate space in a data-driven way. The greater the burden of prototyping or synthesis, the more this “narrowing in advance” drives the cost-effectiveness of development.

• Materials axis — rather than testing AI-proposed material candidates as they come, we filter them at the very top of the pipeline down to “those we can actually make,” aiming for a form that spins the loop between design (information) and verification (sintering) quickly.
• Bio axis — we organise the design of new modalities on a mechanism basis, assembling candidates while cross-checking them against existing knowledge. This is not a basis for asserting efficacy; it is a basis for finding candidates worth advancing to verification and shaping them so they are easy to propose.

Recent focus

On the materials axis, in June 2026 we announced an in-house alloy MI (materials-informatics) search system that filters the candidates proposed by AI, at the very top of the pipeline, down to “only those we can actually make on our own equipment,” and confirmed its operation across several materials systems including space applications. Designing smartly with AI and verifying quickly with sintering — building a base that spins this loop fast is one concrete outcome of SMI.

On the bio axis, we are advancing a mechanism for designing and proposing new modalities, on a mechanism basis, towards solving health challenges. Both axes sit under SMI’s shared aim of organising development results as information and accelerating verification and proposal.