Ulyssean builds integrated hardware and software to secure the data center infrastructure where frontier AI models are trained and deployed, protecting AI model weights against state-sponsored and intelligence-grade threats.
Ulyssean builds integrated hardware and software to secure the data center infrastructure where frontier AI models are trained and deployed, protecting AI model weights against state-sponsored and intelligence-grade threats.
People
Updated 05/18/26Founder
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26Ulyssean PBC is a public benefit corporation building security infrastructure for AGI data centers. Founded by Jacob Lagerros, the company develops integrated hardware and software solutions to protect the facilities where frontier AI models are trained and deployed. The company's work is defined by three core theses: first, that AI data centers will be the most important locations on Earth; second, that the current security posture of these data centers is not ready for AGI; and third, that now is a historical moment to solve this before the majority of future AGI compute infrastructure is built. Ulyssean operates as a vertically integrated team doing hardware security research and engineering, with work spanning decapping accelerators, improving open-source FPGA tools, and formally verifying their security stack. Their technical focus areas include ML, cryptography, FPGAs, formal verification, and side-channel analysis. Jacob Lagerros has presented research on how AI models might exfiltrate data from airgapped environments by modulating their own power signature, or break network topologies by modulating memory accesses to create makeshift radios. The company's first major project is a design partnership with Fluidstack, constructing Security Level 4 facilities hardened against state-sponsored and intelligence-grade threats, built to protect AI model weights representing years of research and billions of dollars in compute investment. Ulyssean aims to scale to every major US-allied data center in the world by 2028. Founder Jacob Lagerros brings deep roots in the AI safety ecosystem. He was previously a researcher at Oxford's Future of Humanity Institute, co-founder of Lightcone Infrastructure, and worked with Metaculus on forecasting. He co-leads the UK Secure Cluster and has presented to the UK National Security Council on AI export controls. He also serves as an advisor to the AI Security Forum and organized the Paris AI Security Forum 2025. Ulyssean is backed by angel investors from Anthropic, DeepMind, Meta, and CrowdStrike, and received $186,000 in funding from the Survival and Flourishing Fund's 2025 S-Process round for general support.
Theory of Change
Updated 05/18/26Ulyssean believes AI data centers will become the most important and most targeted locations on Earth, yet their current security posture is inadequate for the AGI era. By building integrated hardware and software security solutions now, before the majority of future AGI compute infrastructure is constructed, Ulyssean aims to embed robust security into the foundation of these critical facilities. Protecting AI model weights from theft by state-sponsored actors and preventing unauthorized data exfiltration from airgapped environments directly reduces the risk of advanced AI capabilities falling into the hands of malicious actors, thereby mitigating catastrophic and existential risks from AI misuse.
Grants Received
Updated 05/18/26Projects– no linked projects
Updated 05/18/26Discussion
No comments yet. Be the first to share your thoughts.