Most enterprise AI programs don't fail because of the model. They fail because of the infrastructure beneath it: the governance gaps, security blind spots and workflow friction that keep AI locked in pilot mode instead of delivering production-scale business value.
But accessing the agentic enterprise requires far more than just better models. AI agents need a trusted foundation: a single source of enterprise truth, built-in security capabilities, identity-aware access controls and policy guardrails that allow them to operate reliably across business workflows. Without that foundation, even the most capable models cannot safely take action.
At Snowflake, we've long believed there is no AI strategy without a governed data strategy. As enterprises move beyond experimentation and toward deploying AI agents in production, a new infrastructure layer is emerging between foundation models and business applications. This layer is becoming one of the most important investment opportunities in enterprise AI, enabling organizations to operationalize AI securely, govern it consistently and integrate it into the workflows where business value is created.
That's why we're excited to highlight Dust and Gray Swan, two Snowflake Ventures portfolio companies helping define this next phase of enterprise AI.
Each company addresses different challenges, from AI security and governance to enterprise agent platforms and model infrastructure. Together, they represent the foundational capabilities organizations need to deploy production-grade, agentic AI with the trust, flexibility and governance controls that enterprises require.
From AI experimentation to enterprise deployment #
The first wave of AI adoption was defined by experimentation. Organizations tested new models, explored use cases and evaluated potential business impact. Now, enterprises are building AI-powered products, deploying agents and embedding intelligence directly into business processes. Success increasingly depends not only on model quality but on the surrounding systems that make AI reliable, secure and useful in production.
There are three capabilities that are becoming foundational:
- Flexible access to the best models for a given task
- Security and governance controls for AI systems operating at scale
- AI-native workflows that enable humans and agents to work together effectively
Dust and Gray Swan are both helping define these categories.
Bringing AI into everyday work with Dust #
Dust is building organizational AI for better collaboration across teams of people and agents. With traction across GTM and operations teams at AI-native startups and established enterprises alike, Dust gives teams a shared foundation of company knowledge, search and Model Context Protocols (MCPs), on top of which they can orchestrate complex, cross-functional work using any model and reusable, governable agents and skills.
By making AI more collaborative, flexible and governable, Dust helps companies move beyond isolated pilots and toward broader organizational adoption. We’re particularly excited about the growing integration between Snowflake and Dust, helping customers build and scale Cortex Agents and Dust agents grounded in trusted enterprise data.
Building trust in AI with Gray Swan #
Gray Swan is building the runtime security and protection layer required for this next generation of AI applications. Its platform helps organizations identify vulnerabilities, strengthen defenses and deploy AI with greater confidence.
As enterprise AI adoption accelerates, trust, governance and security have become foundational requirements for every organization seeking to scale AI responsibly.
Scaling the agentic enterprise #
At Snowflake Ventures, we invest in companies that empower organizations to unlock more value from their enterprise data and accelerate their AI transformation journey. What excites us about Gray Swan and Dust is not only their individual product execution but how they collectively reflect where enterprise software is headed.
The future of enterprise AI requires more than raw model power. It demands a single source of enterprise truth, secure and governed deployment runtime, and seamless integration into the everyday workflows where decisions get made and work gets done.
We’re thrilled to collaborate with the teams at Gray Swan and Dust as they help build the trusted infrastructure powering the agentic enterprise.
Interested in building with Snowflake? Explore how the Snowflake AI Data Cloud helps organizations develop, deploy and scale AI applications with trusted enterprise data, and register here to join our Snowflake for Startups ecosystem.