By Jon DeWald, CEO and Co-Founder
We recently learned that HelixIntel has been accepted into the NVIDIA Inception program, a program for startups building with accelerated computing. It’s a meaningful moment, and I want to share what’s behind it.
The core of what we’re building is simple to describe and hard to do: train an AI model on proprietary data from real buildings, connected to real insurance outcomes, and use it to help insurers understand risk before it becomes a claim.
Loss control is a resource-constrained function by nature. Teams are covering large portfolios with limited time, and even the best programs can only reach so many properties in a given year. What happens between touchpoints has historically been difficult to see. HelixPrevent exists to change that. It gives insurers structured visibility into the maintenance activity that drives their loss experience, along with tools to provide proactive support where it’s needed most.
The data that powers this is what makes it defensible. Years of maintenance workflow data tied to real claims outcomes across thousands of properties form the foundation we are training on, and it is something no generalist platform can replicate.
Being accepted into NVIDIA Inception reflects the seriousness of what we’re building. It gives our team access to infrastructure, resources, and a community of companies pushing AI forward in meaningful ways.
Insurance partners are engaging with this model, and the results are validating the thesis. We’re excited to keep building and to scale what’s working.
Much more to come.
© 2025 NVIDIA. The NVIDIA logo is a trademark and/or registered trademark of NVIDIA Corporation in the U.S. and other countries.
