# Simon Lab roadmap (summary)

Summary of https://simon-lab.org/roadmap/ (Polish: https://simon-lab.org/pl/roadmap/, German: https://simon-lab.org/de/roadmap/).

Simon Lab is in a private beta. The target is the launch, and the end of the beta, at ACHEMA 2027 in Frankfurt (14–18 June 2027), with an offer made together with trusted partners.

What the platform is built toward: from one structured project scope, the project structure, a visual process model with its simulation, vendor-neutral control structures ready for PLC automation engineering, a CAPEX model built on real procurement, and the engineering documentation, all from one source of truth. Simon Lab does not replace engineering judgement.

## Steps, in order
1. **Now: application hardening.** After fast, AI-assisted prototyping sprints: a current technology stack with reproducible checks, one project identity and one session across the applications, AI agents acting only through typed and validated commands (destructive changes need a person's approval), audit records and self-tests.
2. **Next: infrastructure hardening.** AI agents need far more computing power than office computers, because they work much faster than people and run specialised tools (process simulation, strength calculations (FEM) and flow simulations (CFD)). Compute, storage, backups and secure remote access for trusted partners.
3. **Calibration: projects with trusted partners.** Finished artefacts (PDF documentation, 3D models, results that fit the partners' workflows), calibrated against what their clients and engineers expect.
4. **Access for engineers.** Trusted partners - individuals, engineering companies, universities. They are invited to test the platform on their own in the BYOA model (bring your own agent: each works with their own AI agent), through access to the Simon Lab CLI.
5. **Open source and the community.** The battle-tested platform becomes open source, to invite broader engineering collaboration: engineers from around the world improve it with us. Anyone who wants to can also fork it.
6. **Broader access and infrastructure scale-up.** A tool for engineers to invite their team members, or their own client's team, to work in the project; the project stays transparent also for a client's own AI agent.
7. **Physical architecture and reliability.** Compute for a wider group of clients, reliability, uninterrupted access.
8. **ACHEMA 2027 (Frankfurt, 14–18 June 2027): the offer with trusted partners, and the end of the beta.** The platform is open source, so projects move in and out easily (program code is a language every AI agent understands); nothing is gatekept, the data belongs to the user, and anyone can bring their own compute (BYOC) and host it on their own hardware. Trusted partners supervise the AI, validate results and bring engineering insight; results refined by people, visible in the platform or through exports. The platform speeds the work up; people decide. The engineer leaves drafting and drudgery behind and takes on the role of an architect of chemical engineering.
9. **After launch: infrastructure for the AI agent of choice.** Simon Lab as a company becomes an infrastructure provider, delivering the compute that AI agents need to work in this environment, the only residual value in the age of AI.
