Dust raises $40m Series B led by Abstract and Sequoia for multiplayer AI agents

What Dust does

Dust argues that most companies use AI in a single-player way: each employee talks to a separate chatbot or assistant, and what it finds stays in a private chat window. Its answer is what it calls multiplayer AI, in which people and agents work in a shared workspace with common projects, conversations, to-do lists and files, and with governance controls built in.

The platform links to company data in tools such as Slack, Notion and Salesforce so that agents can work with the wider context of an organisation. SiliconANGLE said it connects to more than 100 enterprise data platforms; Dust's website currently refers to more than 70 connectors, so the count varies by source. The product also includes a cloud-hosted environment for processing files and producing documents, and memory that lets agents learn from user preferences over time.

Dust does not build its own models. Customers choose which model powers each agent, and its website lists more than 20 frontier and open-source models from providers including OpenAI, Anthropic, Google and Mistral.

Traction

According to the company, more than 3,000 organisations use Dust, which has seen more than 300,000 agents deployed, 70% weekly active usage across customers and zero customer churn in 2025. Tech.eu and SiliconANGLE both repeated these figures as company claims. Dust did not publish revenue.

Ramtin Naimi, a general partner at Abstract, said AI operators inside companies such as Datadog and 1Password build agents that collaborate across teams and learn from every interaction. Konstantine Buhler, a partner at Sequoia, said zero churn and 70% weekly active usage show the product is not experimental.

Founders

Dust was founded by Gabriel Hubert, chief executive, and Stanislas Polu. According to Tech.eu, they met at Stanford in 2007, co-founded the data analytics company TOTEMS, which Stripe acquired in 2014, and then spent five years at Stripe.

Polu later joined OpenAI as a research engineer and left in September 2022 convinced that models were powerful enough to be economically transformative but under-deployed because the product layer was missing. Hubert became chief product officer at Alan, the report said.

What the money is for

Dust said it will use the round to push three areas at once: agents that learn and improve automatically as they are used, collaboration features that make humans and agents equal contributors with shared access to projects, tools and context, and infrastructure that keeps governance and orchestration predictable at enterprise scale.

Hubert described the shift as century-defining and said that it is only year three. He said what will change how work gets done is a new type of system that gives humans and agents shared, governed access to the same information and capabilities. The sources did not report headcount or the details of Dust's earlier rounds.