Y Combinator and Partners on LinkedIn

Congrats Zaky and Min!! thrilled to get to watch you build the future of pesticides

Ray is betting that marketers will need their own devtools. I think he’s right.

Congratulations Alex! I predict a generational edtech company in the making...

Seems inevitable to me that AI will be operating everything in the critical path to intelligence, including the power plants it runs on.

Congrats to Kara Holinski and Justin Hunter on Confido $55M Series B led by Insight Partners. Confido automates accounting and planning for consumer brands, including divisions of Mars and Unilever. They joined YC S21 as Outfit, a business for independent fitness instructors. When the economics didn’t work, they chose to pivot and defer Demo Day. It took about a year of searching to find Confido. Justin had worked at consumer companies and seen the manual work firsthand. When he cold-emailed Coca-Cola, someone offered to talk for two hours. He thought they meant call back in two hours. They wanted to spend two hours explaining their problems. They closed Confido’s first customer with a Figma before writing code for it. What he learned is that brands outsourced work to pull data from retailer portals to do accounting and forecasting to update their supply forecasts. It was quite messy. For example, one of their first customers who is a public company wasn’t reviewing transactions below $10,000 since it was so much effort. This was the insight that led them to land on this idea, so they build a product to do the work not just another tool. Now Confido lets a small team run a global consumer brand by doing the accounting, forecasting and production planning that today require entire departments. https://lnkd.in/gHQPRHxB

I made a 3d miniature of my hometown in upstate NY: https://avon.town

I made a 3d miniature of my little hometown in upstate NY: https://avon.town

Congrats to Rishi Choudhary 🏰 and Nitish Poddar on Kastle's $24M Series A led by Insight Partners. Their AI agents have processed $2B in transactions. Kastle builds AI employees for banks, starting with mortgage servicing. They collect payments and answer borrowers' questions over the phone. Today they work with 10 of the top 25 mortgage servicers, including Newrez. During YC S24, they got their AI sales product to about $4k MRR, but customers treated it as a nice-to-have. They decided to give up that revenue and focus on a problem customers couldn't leave unsolved just few weeks before demo day, so the clock was ticking... What they did is they flew around the country to mortgage conferences and they learned about a niche problem that I was surprised about. I t turns out 10M Americans were making mortgage payments over the phone each month! with people in call centers handling the calls, so they found a good problem to tackle: build voice agents on taking those payments. They were able to quickly close some of the largest mortgages like Newrez and go live with them. https://lnkd.in/gDUBDpgY

Excited for Dreamscale in the new YC batch! They have a contrarian opinion that inference for robotics is going to move more toward the cloud than the edge, and they are building the infra to make it super performant.

Congrats to Zubin Singh Koticha, Alexis Gauba and Ben Hylak on Raindrop's Series A led by CRV, bringing their total funding to $50M. they detect failures in AI agents, such as tool call failures and decision errors. Companies like Vercel, Clay, and Speak use them. This is a story how they got started as a terminal and now solving the rise and fall of agent civilizations with Simulations When they applied to YC W24, they were building a coding terminal with AI features that were agentic before the models could reliably do it. They eventually found a bigger problem while building the terminal: it was hard to determine if all the AI features worked 100% correctly for their users. They found this to be a more exciting startup idea to focus on. It turns out it's expensive and hard to use an LLM as a judge for all traces. So they built a very efficient classification model that generalizes and lets them monitor all traces. What's interesting is they started to take off after Claude 4.5 made agentic coding possible. Usage grew exponentially as more agents called more tools and got access to production systems, creating more ways to go wrong and raising the cost of mistakes. Now they're running simulations of armies of agents so companies can prevent failures like those in “Rise and Fall of Agent Civilizations,” describing the OpenAI x Hugging Face incident: https://lnkd.in/gEzbTgES Simulations run agents against synthetic copies of the services they use, so teams can test changes before merge and find failures they hadn't thought to write an eval for. It's integrated into CI so teams can run these checks on every code change. https://lnkd.in/gwPbb-Sb

Memorable found a way to optimize memory with embeddings instead of more tokens which is a powerful new way to do memory

Congrats to Rachit Kataria and William Wang at Centralize on their Series A led by NEA! They automatically build org charts of their customers’ accounts to find the people who can move deals that get stuck. Companies like Cognition, Decagon and Exa use them. It’s a story of how two engineers became very good at sales. One of their customers had a Fortune 10 deal that had been stuck for six months. Centralize found a path through the org, and the customer closed an eight-figure deal within weeks of using them. During YC W24, they had a general product where users could chat with their customer data. But one question kept coming back: how do we find the key decision makers for a sale? They leaned into that and built a way to make the relationship context buried in calls and emails legible at scale. As sales, renewals and expansion run through Centralize, it becomes an up-to-date ledger of customer relationships that’s hard for sales teams to replace. https://lnkd.in/g7-aCFBv

Legora is growing total ARR 10x year over year. Their growth is on par with OpenAI and Anthropic.

You need to build and own your own intelligence: your own agent, your own models, your own memory Join us for the YC Own Your Intelligence Hackathon September 27 in San Francisco https://lnkd.in/gFsHG5Dx

A few weeks ago, I visited Heart Aerospace in LA to record an interview with their CEO and Co-founder Anders Forslund ahead of its first flight. Here is their story.

Congrats to Abraham Heifets on launching WonderTx. It's been fascinating to watch Abe build a "lights out" drug discovery company, and their experimental results are pretty exciting...

Introducing Standard Machines: Teaching AI to Design Advanced Chips

Qlo (YC S26) uses AI to run commercial underwriting end-end. Commercial insurance is a massive part of the US economy and today its run by email and excel. No more.

I can't believe that Pocket is doing a $100M revenue run rate... within 6 months of doing YC. This team is unstoppable!

Hell yes watch this You know what’s crazy? Cosmic Robotics has already built and deployed all the tech you see in this video on the largest construction projects on Earth (solar and data centers) as a seed stage company.