Y Combinator F25 startup Multifactor has raised $15 million in a seed funding round aimed at hardening online accounts against emerging threats from AI agents, while allowing both humans and automated systems to share access without exposing underlying credentials.
The round was led by Nexus Venture Partners, with participation from Y Combinator, Taurus Ventures, Honeystone Ventures, Flex Capital, Pioneer Fund, Ritual Capital, Liquid2 Ventures, and several prominent operators and founders across security, AI, and enterprise software. Multifactor describes itself as a post-quantum security platform that provides the safest access to online accounts for both humans and AI agents, relying on cryptographic techniques that it claims are mathematically unbreakable even in a post-quantum environment.
Rather than acting as a traditional password manager, Multifactor converts online accounts into revocable, capability-based access links. The company compares this to the way Google Docs links can be shared with different permission levels. For banking platforms, email, calendars, social media, and enterprise applications, users can generate links that grant read-only or feature-limited access, with optional constraints such as dollar limits for financial operations. These links can be consumed by either human collaborators or AI agents, while underlying passwords and credentials are never exposed.
All actions taken through the platform are cryptographically signed and audited, and access is enforced through a zero trust, capability-based model where agents can see and modify only what the user has explicitly granted. The company positions this design as a direct response to new failure modes in AI-heavy environments, such as prompt injection, password theft, confused-deputy problems, and cross-agent hijacking, which can arise when autonomous or semi-autonomous agents are given broad access to sensitive accounts.
Multifactor was founded by CEO Vivek Nair, a PhD computer scientist and former CIA officer, and Colin Roberts, a PhD mathematician and cryptographer. According to the company, the platform is based on roughly a decade of cybersecurity research and experience in high-risk national security environments. The service is already publicly available as a free account manager, accessible at the company’s website, and the seed funding is intended to support product development and go-to-market efforts as organizations experiment with agentic AI in production workflows.
Multifactor is an early example of a security product category emerging specifically for the agentic web, where AI agents are allowed to act on behalf of users across multiple SaaS and financial platforms. The model decouples the concept of having the password from having access, and it offers strong auditability and revocability around what agents can do. That may resonate with organizations that want to leverage agents for operational tasks such as reconciliations, monitoring, or customer support, while avoiding direct credential sharing or uncontrolled OAuth flows. It also raises design questions for identity and access management programs, which will need to decide how to integrate capability-based agent permissions and cryptographic audit trails into existing governance and compliance frameworks.
—
By the ID Tech Editorial Team
