Dust raises $40M to push enterprise AI past the single-player era

The Paris- and San Francisco-based platform’s Series B is co-led by Abstract and Sequoia, with Snowflake and Datadog participating. Total funding is now north of $60m.

Dust, the Paris- and San Francisco-based enterprise AI platform, has raised a $40m Series B co-led by Abstract and Sequoia, with participation from Snowflake and Datadog, the company said on Monday.

The round takes Dust’s total funding to more than $60m and follows a $16m Series A in June 2024, also led by Sequoia.

Dust is selling a particular argument about what enterprise AI is missing: that the dominant product shape so far has been an assistant per person, with the context from each session disappearing back into a private chat window once it ends.

The company calls that ‘single-player AI’, and frames its own product as the ‘multiplayer’ alternative: a shared workspace where agents and employees draw from the same projects, conversations, files, notifications, and to-do lists, governed centrally and connected to the systems the company already runs on.

‘What will transform the way we work isn’t the next best model or assistant,’ said Gabriel Hubert, Dust’s co-founder and chief executive, in a statement.

‘It’s going to be a completely new type of system that gives humans and agents shared, governed access to the same information and capabilities so they become true collaborators.’

The framing is positioning as much as product: an attempt to draw a category line between Dust and the wave of single-user copilots from foundation-model labs and software incumbents.

On the company’s own numbers, Dust is now used by more than 3,000 organisations, reached 41,000 monthly active users in April, and has more than 300,000 agents deployed across its platform. The company reports 70% weekly active usage across its customer base and zero customer churn in 2025.

The platform connects to over 100 data sources, layers in memory and agent analytics, and ships SOC 2 Type II certification and GDPR compliance with EU and US data residency, with a contractual commitment from major providers not to train on customer data.

Customer datapoints sit underneath the marketing line. At Vanta, a 46-person revenue team estimates 400-plus hours saved a week, per CRO Stevie Case.

Watershed cut a recurring data-mapping workflow from two to three hours to a few minutes at a 78% success rate. In Europe, Qonto’s case study with Dust puts savings at around 50,000 hours a year across 50-plus specialised agents and 1,000-plus daily users.

The category is getting crowded. Anthropic shipped ten financial-services agent templates inside Claude earlier this month, and Google, Microsoft, and OpenAI have all been pushing variants of agentic enterprise tooling.

Sequoia’s Konstantine Buhler framed Dust’s bet as orthogonal: ‘Most enterprise AI today is single-player: one person, one prompt, no compounding.’ Abstract’s Ramtin Naimi described AI Operators inside customers like Datadog and 1Password as already ‘rewiring how the entire company works’.

There is also a labour-market subtext. Where companies like Klarna have leaned into AI as a hiring substitute, Dust is selling employers a tool that explicitly assumes the workforce stays in place and gets leverage from agents rather than being displaced by them.

The ‘AI Operator’ role, which the company describes as an internal builder inside Ops, Support, Marketing or Sales who configures and runs agent fleets, is the staffing model implied by the product.

Dust was founded in February 2023 by Hubert and Stanislas Polu, who met at Stanford in 2007 and previously sold data analytics company TOTEMS to Stripe in 2014.

Polu went on to a stint as a research engineer at OpenAI, working on mathematical reasoning under Ilya Sutskever; Hubert was chief product officer at French health-tech Alan. The Paris-incorporated company has more recently scaled its US operations out of San Francisco (per Sequoia’s portfolio page).

The Series B will push three things at once, on Dust’s own brief: agents that improve as they are used, collaboration primitives that make humans and agents bidirectional co-contributors, and the orchestration and governance plumbing for enterprise scale. Run-rate revenue was not disclosed.