SCHINES · Project

Brain-Like Chip Wiring That Cuts Power Use and Boosts Signal Routing

digitalPrototypeTRL 3Thin data (2/5)

schines.cnrs.fr

Imagine trying to wire a city where every building needs to talk to every other building — that's the interconnect problem in brain-like computer chips. Current chip designs simply can't handle that many crossing signals without massive power waste and interference. SCHINES found a way to route electronic signals through special materials where electrons behave like light passing through lenses — they can cross paths without interfering, much like two flashlight beams passing through each other. The team at IBM Research and partners built the first switchable devices using these "topological metals" to prove the concept works.

By the numbers

consortium partners across 3 countries

total project deliverables completed

33%

industry participation ratio in consortium

demo deliverable with working gated devices

The business problem

What needed solving

As chips get more complex and AI workloads explode, the wiring inside processors cannot keep up — signals interfere, power consumption skyrockets, and brain-like computing architectures remain impossible to build at scale. Current interconnect technology is the single biggest bottleneck preventing neuromorphic chips from reaching their potential, and no conventional solution can fix it.

The solution

What was built

The team fabricated the first gate-switchable devices using topological metals that enable lossless electronic signal crossing. Across 9 deliverables, they demonstrated that electrons in these special materials can be steered like light through lenses — crossing paths without interference and maintaining signal integrity through topological protection.

Audience

Who needs this

Semiconductor companies designing next-generation chip interconnectsAI chip startups building neuromorphic computing hardwareData center operators seeking energy-efficient computing solutionsDefense and aerospace firms needing ultra-reliable signal routingQuantum computing companies exploring topological protection for qubits

Business applications

Who can put this to work

Semiconductor Manufacturing

enterprise

Target: Chip design companies working on next-generation interconnect architectures

If you are a semiconductor company struggling with signal routing bottlenecks as chip complexity grows — this project demonstrated devices using topological metals that allow lossless signal crossing. With 9 deliverables completed over 4.5 years and IBM Research as coordinator, the results could inform your next-generation interconnect designs where conventional wiring simply cannot scale.

AI Hardware

enterprise

Target: Companies developing neuromorphic computing chips and accelerators

If you are building dedicated AI hardware and hitting power consumption walls with dense neural network architectures — SCHINES demonstrated a design principle for scalable interconnectivity in physical neural networks. Their devices with gate-switchable signal control address the core challenge that existing neuromorphic architectures cannot achieve brain-like connectivity without prohibitive energy costs.

Data Center Operations

enterprise

Target: Hyperscale data center operators and cloud infrastructure providers

If you are a data center operator watching energy costs climb as AI workloads demand more from your infrastructure — neuromorphic chips built on SCHINES principles could eventually deliver brain-like processing at a fraction of current power draw. The project's 3-country consortium including IBM Research produced working gated devices, moving this from theory toward future hardware you could deploy.

Frequently asked

Quick answers

What would this technology cost to license or integrate?

No pricing or licensing information is available from the project data. As a FET Open research project coordinated by IBM Research, commercialization terms would need to be negotiated directly. The technology is at an early stage, so costs would likely involve co-development investment rather than off-the-shelf licensing.

Can this scale to industrial chip manufacturing?

The project demonstrated the first gate-switchable devices as proof of concept. Scaling to industrial CMOS-compatible manufacturing would require significant additional development. The design principles are described as 'highly transferrable' across strained materials, magnetic domains, and heterostructures, which suggests multiple paths to manufacturability.

Who owns the intellectual property?

IP would be shared among the 3 consortium partners across Switzerland, Germany, and France under Horizon 2020 grant rules. IBM Research GMBH as coordinator likely holds key patents. Specific licensing terms would need to be discussed with the consortium.

How does this compare to existing neuromorphic chip solutions?

Based on the project objectives, existing physical neural network architectures cannot achieve the interconnectivity needed for brain-like computing. SCHINES addresses this specific gap using topological metals for lossless signal crossing — a fundamentally different approach from conventional interconnect methods used by current neuromorphic platforms.

What is the timeline to a commercial product?

The project ran from 2019 to 2023 and produced working gated devices. However, as a FET Open project (frontier research), commercial deployment is likely years away. Based on available project data, the next steps would involve scaling device fabrication and integrating with existing chip architectures.

Is there regulatory risk with this technology?

Based on available project data, no specific regulatory hurdles are identified. The technology involves materials science and chip design, which fall under standard semiconductor industry regulations. Export controls on advanced chip technology may apply depending on the application and destination market.

Consortium

Who built it

The SCHINES consortium is compact but heavyweight: 3 partners across Switzerland, Germany, and France, led by IBM Research — one of the world's top corporate R&D labs. With 1 industry partner (IBM) and 2 research organizations, the 33% industry ratio signals that this isn't purely academic work. The absence of SMEs and universities means this was driven by established research powerhouses with direct industry relevance. For a business looking to engage, IBM Research as coordinator provides a credible entry point with deep commercialization experience, though the small consortium size means fewer potential collaboration angles.

IBM RESEARCH GMBHCoordinator · CH
MAX-PLANCK-GESELLSCHAFT ZUR FORDERUNG DER WISSENSCHAFTEN EVparticipant · DE
CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE CNRSparticipant · FR

How to reach the team

IBM Research GMBH in Switzerland — reach out to their neuromorphic computing or advanced interconnect research groups

Order a report on this consortium See IBM RESEARCH GMBH profile →

Next steps

Talk to the team behind this work.

Want an introduction to the SCHINES team to explore licensing or co-development? SciTransfer can arrange a targeted meeting with the right technical contacts.

Order a report on this topic More in Digital & ICT