Quantum

SuperMatics Uses AI to Turn Superconductor Discoveries Into Usable Wire

Read full article at source ↗

Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 04/11/2026, 22:49:25 UTC.

Briefings

Berkeley-based SuperMatics is developing AI-driven systems to convert theoretical superconductors into affordable manufacturing wire, addressing the industry's reliance on trial-and-error. The company uses multimodal foundation models and agents to predict properties, synthesize recipes, and simulate devices, incorporating experimental data into continuous model training. This approach aims to solve the cost and scarcity barriers preventing widespread adoption of high-temperature superconductors in data centers, fusion, and power delivery. The company plans to present its approach at the Plug and Play Silicon Valley November Summit on November 5, 2026, in Sunnyvale, California.

Key points

• SuperMatics employs multimodal foundation models and AI agents to predict material properties, write synthesis recipes, and simulate finished wires or devices for superconductors.
• The company is in talks with global materials and technology companies for joint development programs on existing high-temperature superconductors while building its own new materials pipeline.

Context

SuperMatics is developing AI exclusively for superconductors to bridge the gap between theoretical discoveries and commercial wire. The company targets applications including data center power delivery, quantum computing, and fusion magnets. It is in talks with global materials and technology companies about joint development programs and will present its approach Nov. 5 in Sunnyvale, California.

Limitations

Evidence is limited to the provided press release. Specific dates for joint programs are not confirmed beyond the November 5 presentation. The company's internal training data and specific simulation results are not disclosed.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 04/11/2026, 23:05:27 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeif5nazan2bqbphsoo43dp5rsrm7xildw4virv4mwh5amkluzzb7ku