Datalinx AI quietly turned on the lights in a room most enterprises pretend is clean. New York City and Austin based, founded in March 2025, the company just raised a $4.2 million seed round to deal with the part of AI no one likes to talk about. The data before the demo. The data before the promise. The data before the slide deck lies start to sweat.
Joe Luchs, Founder and Chief Executive Officer, has lived inside this problem at scale. Amazon. Oracle. Beeswax. Billions in contract value moving between AWS and Amazon Ads. When you spend that long watching smart teams stall because their data cannot stand up straight, you either accept it or you build something sharper. Datalinx is the latter, not a bandage but a refinery.
Nicole Landis Ferragonio, Founder and Chief Product Officer, brings the scars and the receipts. A decade inside Amazon Ads, a U.S. patent for privacy safe analytics, and hands on work building Amazon Marketing Cloud. She knows where the bodies are buried because she helped dig the tunnels. Product here is not theory. It is memory.
Jeff Collins, Founder and Chief Technology Officer, adds industrial weight. Former Senior Vice President at Unity, architecting systems that supported more than two billion players. Before that, Intuit, platforms, patents, scale that breaks polite software. Alek Liskov, Founder and Chief AI Officer, rounds it out with deep enterprise data and AI leadership from Intuit and Verizon, managing hundreds across product and engineering. This is not a starter kit team.
High Alpha led the oversubscribed round, joined by Databricks Ventures and Aperiam, with angels like Frederic Kerrest, Ari Paparo, and Arup Banerjee leaning in. Databricks did not just write a check. They pulled Datalinx into the inaugural AI Accelerator cohort in September 2025, an early signal that the plumbing mattered more than the paint.
Datalinx calls itself an AI data refinery, and that word earns its keep. Agentic systems that discover, clean, validate, govern, and activate data where it already lives. No copying. No shortcuts. Cloud native, privacy forward, built for marketing, advertising, and commerce teams tired of explaining why models fail when the inputs lie.
Sallie Mae is already co-developing with them. Early results point to faster go to market, not because the models got smarter, but because the data stopped tripping over itself. Research cited by the company shows 85 percent of AI projects fail at the data layer. Datalinx is betting that fixing the link fixes the chain.
This $4.2 million is not fuel for noise. It is pressure applied to a bottleneck every enterprise knows but avoids. When AI moves from experiment to expectation, the companies that win will not be the loudest. They will be the cleanest. And Datalinx AI is building for that moment, one stubborn dataset at a time.
