With global supply chains under strain from tariffs, trade disputes and decades of underinvestment, a San Francisco startup is betting artificial intelligence can help reinvent how chemicals are discovered and manufactured.
Mstack, a vertically integrated specialty chemicals company, announced on September 24 the launch of Chemstack AI, a closed-loop platform that the company says can compress chemical synthesis and commercialization timelines from roughly 18 months to a matter of days. The rollout coincides with what the company describes as explosive growth — a tenfold jump in revenue and a customer base that now spans more than 100 enterprises across North America, Asia and the Middle East.
“We’re not just another chemical company — we’re fundamentally redefining what’s possible when AI becomes the world’s smartest chemist,” said Shreyans Chopra, Mstack’s founder in a press statement. “Our 10x growth trajectory demonstrates that the market desperately needs innovation-driven supply diversification, and traditional players simply can’t innovate synthesis routes at the speed and scale necessary to create viable alternatives.”
The company’s model is gaining traction at a critical moment. Rising tariffs in the United States — ranging from 10 to 50 percent on certain chemical intermediates — and growing geopolitical rifts have forced major buyers, including agrochemical producers, to seek alternatives to Chinese suppliers. At the same time, legacy firms remain tied to decades-old production methods that can take years to adapt, leaving entire sectors exposed to bottlenecks.
Mstack’s answer is Chemstack AI, which it describes as the chemical industry’s first “closed-loop” AI ecosystem. Its three core tools include LiteratureIQ, which translates scientific papers into searchable knowledge graphs; RetroRank, an AI engine that proposes and ranks synthesis routes with reported accuracy levels exceeding industry standards; and Experimentation Assist, which uses Bayesian optimization to test reactions in the lab, creating a feedback loop between algorithms and human researchers.
The system, the company argues, does more than automate individual tasks: it mimics the intuition of a chemist, transforming what has historically been a linear, manual process into what Mstack calls a “self-correcting discovery engine.”
Since securing funding from Lightspeed and Alphawave, the San Francisco-based firm has expanded rapidly. It has opened an R&D lab in Hyderabad staffed by more than 25 scientists, developed custom molecules across categories from coatings and adhesives to oil and gas, and built flexible supply chains that span the United States, India, South Korea and the Middle East.
The company points to its traction in India’s agrochemical market as proof of its approach: AI-optimized synthesis routes, it says, have enabled more predictable local supply, better product consistency and lower costs.
Mstack’s ambitions extend well beyond its current offerings. The company is now building what it calls a “Chemistry World Model,” a platform designed to make advanced discovery capabilities available not just to pharmaceutical giants but to organizations of all sizes.
For an industry where knowledge remains concentrated among a handful of global players, the promise of democratizing chemical innovation may prove transformative, if Mstack can deliver on its claims.