A pair of teenage innovators walked into Paul Graham’s backyard with an idea few in the agriculture world believed in — an AI system designed to create better pesticides. By the time they stepped out, they had a refreshed business strategy, a new company identity, and eventually, Graham’s personal investment. That company is now known as Bindwell, and it has quickly become one of the most promising startups in agricultural biotechnology.
Bindwell Raises $6 Million in Seed Funding
Bindwell has successfully closed a $6 million seed round, jointly led by General Catalyst and A Capital, with Paul Graham participating as an individual investor. Instead of offering AI tools to long-standing agrochemical giants, the company is now using its own models to design original pesticide molecules and directly license the intellectual property. This shift reflects a bold attempt to modernize a sector that continues to depend on decades-old chemical formulas.
Why AI-Driven Molecules Matter
Over the last 30 years, pesticide usage has doubled, yet up to 40% of global crops are still destroyed annually by pests and diseases, according to the UN. As insects adapt and develop resistance, farmers rely on increasingly stronger chemicals — a cycle that harms ecosystems, raises costs, and accelerates resistance.
Bindwell believes AI can break this loop by creating new, precise, and targeted molecules built specifically for today’s agricultural threats.
Teenage Founders With Real-World Motivation
Bindwell was launched in 2024 by Tyler Rose (18) and Navvye Anand (19), who adapted AI drug-discovery techniques for modern agriculture. Their initial collaboration began at the Wolfram Summer Research Program in 2023, where they built PLAPT, a model later referenced in a Nature Scientific Reports study.
Their interest in pesticides came from personal observation:
- Rose learned about farming challenges from his aunt in China.
- Anand, with roots in Punjab, saw how limited pesticide choices impacted crop yields.
A Pivot Inspired by Paul Graham
The founders joined Y Combinator’s Winter 2025 batch planning to sell AI tools to pesticide companies. But most legacy firms were hesitant to incorporate AI into discovery workflows. During a visit to Paul Graham’s home, he recommended a different strategy: use their models themselves to discover molecules instead of selling tools. This advice shaped Bindwell’s entire direction.
Graham later wrote on X that the founders “will probably do alright.”
Inside Bindwell’s AI Suite
Bindwell has built an advanced AI stack focused on reducing model hallucination and ensuring high-confidence predictions. Their tools include:
Key AI Components
- Foldwell – a structural prediction model inspired by AlphaFold, built using a custom diffusion system.
- PLAPT – an open-source protein-ligand model capable of analyzing every synthesized compound in under six hours.
- APPT – a protein-protein interaction system for screening biopesticides, performing 1.7× better than existing benchmark tools.
- Uncertainty Engine – flags predictions that require additional data.
Together, these systems can analyze billions of compounds and operate four times faster than AlphaFold 3.
How Bindwell’s Approach Differs
Current pesticide development relies heavily on trial-and-error testing, where chemists must create and analyze thousands of compounds. Bindwell’s AI shifts the focus to target-based discovery.
The model identifies proteins found only in harmful pests — and absent in beneficial organisms — then designs molecules that bind to those proteins and disrupt their function.
Progress and Future Plans
Bindwell is currently validating its AI-designed molecules in its San Carlos laboratory and collaborating with external partners for further testing. The company is also in early talks with major global agrochemical firms, with its first licensing deal expected soon. They have initiated discussions with groups in India and China to conduct field trials.