The Sphinx has always been a guardian of riddles, daring you to answer or be swallowed whole. Now a different Sphinx has stepped onto the stage in New York City, less sand and stone, more code and conviction, and it just pulled in $9.5 million in Seed funding to change how artificial intelligence engages with structured data. The round was led by Lightspeed Venture Partners, joined by Bessemer Venture Partners, Box Group, K5, and Impatient VC, with extra firepower from Steve Cohen, Naveen Rao, and executives from Databricks, Windsurf, and Together AI.
Artificial intelligence loves to flex in natural language. Ask it to summarize War and Peace or draft a love note and it plays the part. But ask it to grind through the messy depths of enterprise data, to reason and validate across structured information, and suddenly the show falls apart. That disconnect is what Rohan Kodialam and Jamie Bloxham saw from different vantage points, Kodialam leading AI research in the Data Strategies Group at Citadel, Bloxham as an early technology lead at MosaicML and software engineer at Scale AI. They didn’t just notice the gap; they built Sphinx to close it.
The company’s first release, Sphinx Copilot, is embedded directly in the workflows that over 93 million Jupyter notebook users live in. It’s not about auto-completing code and calling it a day. It’s about agentic reasoning for data science, exploring, verifying, and grounding insights in quantitative evidence. That’s why companies like Oats Overnight are already leaning in. CEO Brian Tate put it plainly: his team is uncovering shopper behavior in minutes instead of hours or days. That’s not just faster, it’s a shift in what’s possible when decisions are chained to data.
With $9.5 million in the bank, Sphinx is pressing forward on expanding across consumer packaged goods, retail, and financial services, while eyeing new arenas like healthcare and manufacturing. The team is building what it calls a “talent-dense” roster of researchers and engineers in New York, pushing forward research in representation learning and reinforcement learning focused on structured data. The focus isn’t on chasing generative AI hype; it’s on producing systems that stand up under enterprise scrutiny and deliver insights that move markets.
Backers like Lightspeed’s Bucky Moore don’t show up for smoke and mirrors. They back companies that rewire how industries operate. This Sphinx doesn’t guard riddles, it solves them, at the speed and scale the $100 billion data insights market demands.
