Rohan Kodialam on Sphinx AI and building copilots for enterprise workflows

Summary

Rohan Kodialam is co-founder and CEO of Sphinx AI, a startup building AI copilots for data-heavy enterprise workflows. The core pitch is that large language models handle text, code, and images well but struggle with the tabular, semi-structured, and time-series data that underpins supply chains, manufacturing, and high finance. Sphinx targets that gap.

The initial commercial focus is quantitative finance. Kodialam, a former quant team lead at Citadel, says the earliest customers are hedge funds running day-by-day or intraday analysis — firms where a single data insight can generate billions in a year. He declines to name the first customer but confirms the segment is active quant traders rather than longer-horizon fundamental investors. The company runs a hybrid go-to-market: direct enterprise sales alongside a bottom-up motion where data scientists and quant researchers use the product off the shelf.

On the Excel question

Kodialam is direct that Microsoft Copilot in Excel is not a competitive threat worth worrying about. Serious data work — the kind that scales to production and generates material financial returns — runs on a different stack. Over time, he sees Excel becoming an interpretability layer on top of tools like Sphinx rather than a place where real work gets done. Prompt-based interfaces, he argues, are already easier than spreadsheet formulas.

Pricing and value capture

Sphinx charges on usage today. Kodialam acknowledges the value-capture problem candidly: hedge fund clients won't disclose what they found with the tool, making it structurally difficult to price against outcomes. The longer-term ambition is full autonomy — solving large-scale data problems in seconds without human intervention — at which point he expects the value, and the pricing power, to become obvious. For now, usage-based billing keeps the model simple.

Talent positioning

Sphinx has recruited from hedge funds (excluding Citadel, where Kodialam holds a non-solicit agreement) and has won candidates over Anthropic and OpenAI. His framing of the talent market is that foundation lab jobs are a bet on AGI, hedge fund jobs are a hedge against the AI bubble popping, and Sphinx offers something narrower — focused application work for people who want to solve one problem well.