ideas

Most of the physical products around us are mixtures rather than brand-new molecules: paints, adhesives, coatings, lotions, sunscreens, and foods. Right now, ingredient bans like PFAS phase-outs and tighter environmental rules are forcing companies to redo thousands of recipes at once, while AI tools can suggest endless new combinations on screen. The bottleneck is mixing a batch in a jar and measuring how it holds up.

There is a big opportunity to build a robotic formulation foundry with a public price list and an API. Customers send a recipe, robots mix it from a stocked library of commercial raw materials, and automated instruments stream test results back as they finish, from viscosity and pH in minutes to hardness, adhesion, and weathering over days. Instead of every manufacturer running slow, isolated bench tests, one high-throughput lab could turn physical R&D into a fast feedback loop.

Cool companies in this space:

• Albert Invent (AI operating system for chemistry and formulation R&D)
• Uncountable (R&D data platform for industrial labs)
• Periodic Labs (autonomous physical labs for materials discovery)
• Plasmidsaurus (fast, cheap turnaround that unlocked latent demand in sequencing)
• Chemspeed (automated benchtop synthesis and formulation hardware)