FundingNovember 14, 2026 · 3 min read

LangChain raises $300M at a $3B valuation to sell a neutral layer for AI agents

LangChain raised $300 million at a $3 billion valuation in a round led by CapitalG, more than doubling its price from a year ago as it pitches LangSmith as a model-neutral platform to run, improve and govern AI agents.

Two monitors on an office desk at dusk showing an abstract agent workflow graph and code traces, with city lights behind
Illustration: Venture Block

LangChain has raised $300 million in a Series C round at a $3 billion valuation, the San Francisco company said on Friday. Existing investor CapitalG, Alphabet's growth fund, led the round, with significant participation from new investors GIC and DFJ Growth and from existing backers IVP and Sapphire Ventures.

The price is more than double the $1.25 billion valuation LangChain carried when it raised $125 million in October 2025 in a round led by IVP. CapitalG and Sapphire first came in on that Series B. Salesforce Ventures, Insight Venture Partners, QuantumLight and MongoDB Ventures joined this time, and Sequoia, Amplify and Benchmark put in more money, according to the company.

From open-source library to agent platform

LangChain started in late 2022 as a side project by Harrison Chase, now chief executive, and became a company in early 2023 with co-founder Ankush Gola, now chief technology officer. Its original open-source library helped developers wire early language models to web search, APIs and databases, tasks the models handled poorly on their own. As frontier labs built those abilities into their own tooling, TechCrunch notes, LangChain shifted its business toward running and managing AI agents.

The commercial product is LangSmith, which the company describes as a framework-agnostic platform to run, improve and govern agents. LangChain now groups it into three parts:

The pitch is neutrality. Chase frames it as letting every company "own its intelligence": pick whichever model suits a task, keep control of data and infrastructure, and keep improving its agents. In its investment note, CapitalG says LangSmith works with any model and runs in LangChain's cloud, self-hosted, or inside a customer's own AWS, Azure or Google Cloud account.

Customers and usage

LangChain says more than 7,000 organizations use LangSmith, including startups such as Harvey, Rippling and Listen Labs and large companies including Nvidia, Toyota and Salesforce. Two customer figures come from the company: Clay monitors hundreds of millions of agent runs a month on the platform, and Workday cut the time to deploy a new agent from five months to under a day.

The open-source side has grown too. CapitalG says LangChain's open-source projects have passed 1 billion total downloads and now see more than 350 million downloads a month, including Deep Agents. At the Series B, the two core packages, LangChain and LangGraph, had about 90 million combined monthly downloads. CapitalG says many of LangChain's largest enterprise deployments started with a developer who already knew the frameworks.

Some of the new money comes from inside the customer base. Salesforce is listed as a LangSmith user and now invests through Salesforce Ventures, which says it had known the team for several years before joining this round.

Crowded market, neutral pitch

LangChain is selling into a crowded field. CapitalG acknowledges that the model labs and hyperscalers are all shipping their own agent tooling, and calls much of it very good. Its case is that an enterprise running hundreds of agents across several model providers and clouds still wants one layer it controls to build, evaluate and govern them. The fund also points to inference costs: teams that can swap models without rebuilding agents, and prove with evaluations that nothing broke, can move to cheaper open-weight models where cost matters more.

Money keeps flowing toward the infrastructure around agents. This week Venture Block covered Mercor's $750 million raise to sell AI evaluations to enterprises and Mirendil's $1 billion Series A for AI research agents. Earlier this month, Meta bought Conductor, a tool for running many coding agents in parallel.

LangChain plans to show new LangSmith capabilities at a release event on Friday, November 20.

Newsletter

The Block Brief

The stories that matter in startups and venture, in your inbox. Free.

Check your inbox to confirm.

Something went wrong. Try again.

More in Funding