2027 Research Roadmap.
One new research area: chip design.
Where our research stands at the end of 2026, said as what can be checked today, and the one new research area we plan for 2027.


2026: what we delivered.
Each item below is live on this site today, with the evidence behind it.
MS Duplex introduced
Self-improving agents for science, research and engineering, connected to the AI teams already use over MCP. Available as a license and through a pilot on your own work.
MS Lattice across twenty-three domains
Physics-based inverse design for materials discovery and optimization, in twenty-three material domains and four modes, run by your own AI agent over MCP.
Blind tests on record
In computational tests, MS Lattice found both sealed target ceramics without being shown them and rebuilt four compounds it had never seen, with every result recorded.
Two research directions
Neuro-Analytic AI and Adaptive Learning, Searching and Optimization, each set out as a research direction with a product built on it.
2027: chip design.
Chip design is the one new research area on our 2027 roadmap. We plan to research how to find chip design candidates worth further validation under targets such as performance, power consumption and area. It is planned research, so there is no product, release date or performance figure to share yet.
Targets set by the designer
Performance, power consumption and area, weighed together as the requirements and preferences of a design.
Candidates worth validating
The aim is a set of design candidates that merit the slower, more expensive validation that follows.
The same research direction
The work builds on Adaptive Learning, Searching and Optimization, the direction behind MS Lattice.
MS Duplex and MS Lattice keep developing alongside it.
Follow the research.
News on our research and our products appears on the blog first. If your team has a problem in one of these areas, talk to us.

