Bindwell

Building Better Pesticides with AI

April 20, 2025

Good morning and Happy Easter! 🐇

It’s Sunday, April 20th — may your baskets be full, your chocolate be dark, and your inbox be blessed with a solid read.

Let’s get into it.

While most people use AI to write emails, this week’s company is using it to kill bugs (scientifically, of course).

Bindwell is using AI to develop better pesticides. Ones that protect crops and the planet. The goal? Faster discovery, more targeted spraying, and less harm to the environment.

As always, if you have any feedback or ideas on how to improve the newsletter, just hit reply - I’d love to hear from you!

First time reading? Sign up here

Know someone who invests in or builds ag businesses?
Share this email with them!

Ag Breakdowns covers the companies revolutionizing the agriculture value chain. If that’s you, apply here

THE BIG PICTURE

Market outlook

The pesticides market is set to double by 2032, from $45.7B to $92.6B. (Source)

Biopesticides (eco-friendly alternatives) are growing fast: $7.72B in 2024, on track to hit $15.66B by 2029. (Source)

AI in agriculture is catching up too, headed for $4.7B by 2028. (Source)

Consumers and regulators want cleaner solutions, opening up a potential $305B agrochemical opportunity. (Source)

Other players in the space

What’s coming next?

Better results: AI could double the impact of biopesticides by 2029 through smarter discovery and delivery.
Climate-ready tools: Predictive AI will help farmers prepare for heatwaves, droughts, and floods.
Stronger defenses: Expect more investment in cyber protection and insurance for farm tech.
Wider access: If pricing is fair, AI tools could reach small farms around the world, not just the big players.

THE ESSENCE

How does it work?

Bindwell built an AI platform that speeds up the entire pesticide discovery process. It helps researchers find effective, targeted pesticides faster and cheaper by using three powerful tools:

Foldwell
Think of this as their protein-mapping engine. It figures out the structure of pest proteins 4x faster than AlphaFold 3. This is the first step to designing a pesticide that hits the right target.
PLAPT
This tool checks how different molecules interact with those pest proteins. It can scan 177 million known compounds in just 6 hours (that would take 242 years using old-school methods). Even your laptop can run it. It processes 700,000 molecules per second.
APPT
This model helps identify biopesticides by predicting how proteins interact with each other. It’s 1.7x more accurate than other tools in spotting the right biological agents.

The problems they’re solving

Current pesticides just aren’t working well anymore.
Despite using more, farmers still lose up to 40% of crops to pests. Resistance is a huge issue.
No real innovation.
Only 40 new active ingredients have been released since 2010. Most are just slightly tweaked versions of old ones.
Development is still stuck in the past.
While pharma has gone all-in on AI, pesticide R&D is still slow and expensive. $300M and 11 years per product.
Lack of precision.
Today’s pesticides can harm bees, pollinators, and soil organisms. That’s bad for the environment and long-term farming.

The products

A full AI platform for pesticide discovery
Foldwell (protein prediction)
PLAPT (ultra-fast molecule screening)
APPT (biopesticide discovery)
No-code tools so anyone can use the tech
Open-source software for ag bioinformatics

Unique value proposition

Speed: Scan and discover new compounds in hours, not years
Precision: Only targets harmful pests, sparing beneficial insects
Lower cost: R&D costs drop by ~80%, and products hit the market 50% faster
Access for all: Even small ag businesses can use their no-code tools
Environmentally friendly: Could cut pesticide use by up to 90%
Collaborative: Open-source tools encourage more innovation
Fast response: Can adapt quickly to new pest threats

THE BUSINESS

Industry: Pesticide Development
Headquarters: San Francisco, CA
Year founded: 2024
Team size: 3
Funding: $500k in pre-seed investment (March 12, 2025)
Investors: Character(VC), Lombardstreet Ventures, Y Combinator
Business model: AI-powered pesticide discovery tools for the agrochemical industry. They offer a platform with three proprietary AI models that accelerate pesticide development.
Traction: Backed by Y Combinator, Character, and Paul Graham. They've developed three breakthrough AI models, including Foldwell (4x faster than AlphaFold 3) and PLAPT, which can screen 177 million compounds in hours versus years with traditional methods.
Website: bindwell.ai
Socials: LinkedIn

Founder profile

Bindwell was started by Navvye Anand and Tyler Rose, two engineers who met at the Wolfram Summer Research Program in June 2023.

Navvye is from India and studied at Caltech; Tyler is from China and worked at Wolfram Research.

Both grew up visiting farms in their home countries, a shared experience that sparked their drive to improve agriculture through technology.

Opportunities

No opportunities at the moment, but you can reach out HERE

Want to learn more?

Introducing APPT: Affinity Protein-Protein Transformers. (Bindwell)
Bindwell (YCombinator)
Foldwell Comparison Video (Youtube)

THE TAKEAWAY

Great things

Blazing speed: Their platform processes molecules at an insane rate. What used to take decades can now be done in one workday.
Better for the planet: Their tech makes highly specific pesticides that don’t mess with helpful bugs or the soil. A massive win for sustainable farming.

This may cause problems

Regulation takes time.
No matter how fast the tech is, getting EPA approval still takes years.
Big players aren’t easy to move.
Companies like Bayer and Syngenta already dominate. Bindwell will need to break through industry inertia.
Data limits exist.
Their models are strong, but they’re only as good as the data they’re trained on. In rare pest cases or new regions, results may vary. Real-world validation also takes time.

What do you think of this week's company?

Hey👋

Thanks for reading!

Want to help me improve?
Here are some questions for you:

Was this useful?
What did you not like?
How would you improve this newsletter?