Dust raised $40M from Sequoia, Snowflake, and Datadog to build a multiplayer AI platform that gives enterprise teams a shared memory layer across assistants.
Enterprise AI has a memory problem. When a salesperson spends an hour prompting a chatbot to research a client account, that work vanishes the moment the chat window closes. The next day, a solutions engineer runs the same queries from scratch. Multiply that across thousands of employees and the productivity promise of artificial intelligence starts to look like a leaky bucket.
Paris-based startup Dust, legally incorporated as Permutation Labs SAS, announced a $40 million Series B on Monday to fix exactly that. The round was led by Abstract and Sequoia Capital, with strategic participation from Snowflake and Datadog, bringing the company's total funding past $60 million, according to SiliconAngle.
The framing Dust uses is deliberately simple: today's enterprise AI is "single-player." Each employee runs their own assistant, their own queries, their own context. Dust wants to make it "multiplayer," building infrastructure so that the context and knowledge from one worker's AI session becomes accessible across the organization.
The institutional knowledge problem
Behind Dust's pitch is a structural flaw in how enterprise software was designed. One-to-one interactions were the template for traditional tools, and the current wave of copilots and chatbots inherited that same architecture. Organizational knowledge does not compound: an insight surfaced by one employee's AI agent stays locked inside a private chat history rather than feeding a broader institutional record.
This is not a niche complaint. As companies have piled into AI deployments, the gap between individual productivity gains and organization-wide transformation has become one of the more persistent frustrations in enterprise software. Dust's pitch is that coordination, not raw model capability, is now the binding constraint.
The funding round carries its own signal beyond the dollar figure. Snowflake and Datadog are not passive financial investors; both companies sit at the center of enterprise data infrastructure. Their participation suggests Dust's approach to shared AI context connects directly to where the data layer is heading, and Sequoia's involvement places Dust in a cohort of enterprise bets the firm has made as AI spending shifts from experimentation to full deployment.
The competitive landscape
Dust is not the only company arguing that isolated AI assistants are a structural problem. The broader market has seen a surge of startups positioning around coordination, workflow, and agentic collaboration, even as the largest players, from Microsoft Copilot to Salesforce Agentforce, have built their own takes on enterprise AI integration. Forbes has tracked how the scale of that competition has grown relentlessly.
What sets Dust apart, at least in its own framing, is a focus on the organizational layer rather than the model layer. It is not building a better language model; it is building connective tissue between the models employees already use. Whether that proves a durable moat or a feature hyperscalers eventually absorb is the central question for any middleware bet in this market.
Funding pressure across the sector is rising. NVIDIA was reported this week to be in advanced talks to lead a $20 million round in Simplismart, an Indian AI deployment startup, at a valuation approaching $100 million, up from roughly $25 million in late 2024, per NewsBytesApp. Picks-and-shovels plays beneath the model wars are attracting serious capital.
With Snowflake and Datadog as backers, integrations with data pipelines and observability stacks should accelerate. OpenAI's reported plan to nearly double its workforce to 8,000 by year-end, as CNBC covered in March, underlines how hard the largest players are pushing into enterprise accounts. The window for coordination-layer startups to establish themselves before incumbents move in may be shorter than Dust's backers would prefer to acknowledge.
The history of enterprise middleware is littered with companies that were right about the problem but wrong about the timing. Dust is betting that enterprises, now deep enough into AI rollouts to feel the coordination pain acutely, are finally ready to pay for the fix.
---
FAQ
What is Dust's multiplayer AI platform?
Dust builds infrastructure that lets AI agents and assistants share context and memory across an organization, rather than keeping each interaction siloed in individual chat windows. The goal is to make insights generated by one employee's AI session available to teammates.
Who backed Dust's Series B?
Abstract and Sequoia Capital led the $40 million round. Snowflake and Datadog also participated as strategic investors, bringing Dust's total funding past $60 million.
How does Dust differ from tools like Microsoft Copilot?
Copilot and similar products remain primarily individual productivity tools. Dust targets the coordination layer, making AI-generated insights shareable and persistent across teams rather than locked in private sessions.
Why are Snowflake and Datadog investing in an AI startup?
Both companies are core enterprise data infrastructure providers. Dust's shared-memory architecture connects directly to how enterprise data is stored, queried, and observed, making the investment a natural adjacency for each firm.