SONRAI SECURITY - The New York and New Brunswick firm building a firewall for cloud identity. Logo: company brand asset.
"Permission guardrails prevent unauthorized actions from ever taking place - across agents, humans, and machines."
Most cloud security tools are excellent at one thing: telling you what is wrong. They generate findings, rank them, and hand security teams a very long list. Sonrai Security built its business on a different premise - that the fastest way to shrink a cloud attack surface is not another dashboard, but a control that closes the doors nobody is using.
That idea has a home in a specific, awkward statistic. In a typical enterprise cloud environment, the overwhelming majority of granted permissions are never actually used. Each one is a standing invitation. Sonrai's answer, the Cloud Permissions Firewall, continuously analyzes real permission usage across every identity and automatically blocks the unused ones with native, organization-level cloud policies - handing access back only when someone genuinely needs it.
The company was founded in 2017 by Brendan Hannigan and Sandy Bird, the pair behind Q1 Labs, a security-intelligence firm acquired by IBM. Hannigan went on to run IBM's security division before turning his attention to the problem he could not stop thinking about: identity sprawl in the public cloud, growing faster than any human team could track. Sonrai is his second act, and it is deliberately built for the cloud-identity era rather than retrofitted into it.
Today the company operates from New York City with deep engineering roots in Fredericton, New Brunswick, employs roughly 75 people, and counts enterprises like Sandisk, Fiserv, ISO New England and PG&E among its 2025 customers. It has raised about $88.5 million and, by its own account, closed 2025 with four-times ARR growth.
Sonrai's platform is built on a graph that models the relationships between identities - human, machine, and now AI - the permissions they hold, and the data those permissions can reach across AWS, Azure, Google Cloud and Kubernetes. That structure matters because real attacks rarely exploit a single misconfigured setting. They chain permissions together, escalating and moving laterally until they reach something valuable. If you cannot see the chain, you cannot cut it.
On top of that graph, Sonrai enforces least privilege as an operating default rather than an aspiration. Its CIEM+ capability hunts for "toxic permission chains" - combinations of privileges that together create an escalation path - and disrupts them. The Cloud Permissions Firewall makes the enforcement itself boringly practical: it works at the organization level, deploys in about a day, and gives developers just-in-time access so there is no standing privilege sitting around to be stolen.
Approximate, directional figures that frame why Sonrai exists - the gap between permissions granted and permissions actually used.
"WALLy is the first AI agent designed to fix privilege risk, not just identify it."
Sonrai Security · WALLy launch, October 2025Continuously analyzes real permission usage and automatically blocks unused privileges using cloud-native, org-level policies. Access becomes dynamic and on-demand - no manual intervention, no broken DevOps.
Identifies and disrupts hidden attack paths by analyzing toxic permission combinations across multi-cloud environments, turning entitlement data into concrete escalation-path defense.
An AI security agent that autonomously removes unnecessary privileges, enforces just-in-time access, and blocks risky permissions across humans, machines, AI agents and third parties - using native IAM controls to act safely.
The original graph-based platform that identifies and monitors every relationship between identities and data inside an organization's public cloud - the data model everything else is built on.
Sonrai is a B2B SaaS company. It sells subscriptions to mid-market and enterprise organizations - typically scoped by cloud footprint and identity count - through direct sales and a partner ecosystem that includes the AWS, Google Cloud and Azure marketplaces.
Its customers skew toward regulated, multi-cloud enterprises where identity governance is both hard and consequential. Named 2025 accounts include Sandisk, Fiserv, ISO New England, BAL, PG&E, Basis, Haemonetics, LiveView Technologies and PIB Group. The company reported doubling its enterprise customer base in 2021 and has continued to add logos through its Cloud Permissions Firewall push.
The strategic through-line is friction removal. Security teams and DevOps teams have long fought over the same permissions - one wants them locked down, the other wants them available. By building enforcement into the cloud provider's own native controls, Sonrai aims to satisfy both sides at once, which is a large part of why the model has scaled.
| Round | Amount | Date | Lead / Investors |
|---|---|---|---|
| Series A | $18.5M | 2019 | Polaris Partners, Ten Eleven Ventures |
| Series B | $20M | 2020 | Menlo Ventures, Polaris, Ten Eleven |
| Series C | $50M | Oct 2021 | ISTARI (lead), Menlo, Polaris, Ten Eleven, NBIF |
The $50M Series C, led by cyber-risk firm ISTARI, brought total funding to roughly $88.5M and was earmarked for R&D and global sales expansion. Reported figures are drawn from public announcements; revenue and valuation specifics are estimates or undisclosed.
The cloud security market is crowded. In the CIEM and cloud-identity space, Sonrai competes with the likes of Wiz, Palo Alto Networks' Prisma Cloud, CrowdStrike, Microsoft's Entra Permissions Management (formerly CloudKnox), Tenable's Ermetic, Orca Security and Britive. Many of these tools are strong at surfacing risk.
Sonrai's differentiation is the move from finding to fixing. Rather than stopping at a prioritized list, it uses native cloud policies to actually block unused and risky permissions, then restores them on demand. Its 2025 launch of WALLy pushes that further - an AI agent that acts on privilege risk autonomously, framed explicitly as fixing rather than flagging.
The company also read the identity shift early. As machine identities exploded and AI agents began appearing in production with real permissions, Sonrai extended its model to govern them the same way it governs people. Its current positioning - "we stop rogue agents" - reframes a decades-old least-privilege principle for a very current problem.
Where does it fit in the stack? Alongside, and increasingly inside, the identity layer of cloud security - the place where CIEM, cloud-native PAM and entitlement analysis converge. It is the firewall analogy applied to permissions instead of packets.
Brendan Hannigan and Sandy Bird, former Q1 Labs executives, launch the company across New York City and Fredericton, New Brunswick.
Polaris Partners and Ten Eleven Ventures back the graph-based cloud identity platform.
Menlo Ventures joins to accelerate development of the Sonrai Dig platform.
Total funding reaches ~$88.5M amid 300%+ YoY revenue growth and a doubled enterprise customer base.
The platform is reframed around one-click least privilege and attack-path disruption.
Sonrai introduces the first AI agent to safely fix cloud privilege risk and closes the year with rapid revenue expansion.
Sonrai secures cloud identity and access. Its platform maps every relationship between identities, permissions and data across AWS, Azure, Google Cloud and Kubernetes, then automatically enforces least privilege by blocking unused permissions and disrupting risky attack paths.
It was founded in 2017 by Brendan Hannigan (CEO) and Sandy Bird (CTO), both former executives of Q1 Labs, a security company acquired by IBM.
It is Sonrai's flagship product that continuously analyzes permission usage and automatically blocks unused privileges using cloud-native, org-level policies, delivering just-in-time access without disrupting DevOps.
Roughly $88.5M in total, including a $50M Series C led by ISTARI in October 2021, with earlier rounds backed by Polaris Partners, Ten Eleven Ventures and Menlo Ventures.
WALLy is Sonrai's AI Cloud PAM agent, introduced in 2025 and positioned as the first AI agent that safely fixes cloud privilege risk by autonomously removing unnecessary permissions and enforcing just-in-time access across humans, machines and AI agents.