Fundamental

Fundamental stepped out of stealth in February 2026 with $255M in funding and a message that cuts through the noise of the startup ecosystem with unusual clarity. Built by DeepMind alumni and led by CEO and Co-Founder Jeremy Fraenkel, this is not another layer riding momentum. It is a direct challenge to where intelligence actually lives inside organizations, buried in tables, waiting for someone to ask better questions.

The premise sounds simple until you sit inside it. Enterprises do not run on prompts. They run on structured data. Rows, columns, decades of operational memory stacked in databases that rarely translate into forward-looking decisions. While most of the startup ecosystem sprinted toward language and image models, Fundamental leaned into the neglected surface area and built NEXUS, a Large Tabular Model engineered specifically for structured data. Pre-trained on billions of real enterprise tables, it treats prediction as a primary output, not a side effect.

Point NEXUS at your data and define the outcome. Fraud detection sharpens without guesswork. Predictive maintenance stops being reactive. Demand forecasting shifts from rearview mirror to forward signal. The real play here is not automation. It is compression. The distance between data and decision shrinks until latency becomes a liability competitors cannot afford.

Capital followed conviction. The $225M Series A, led by Oak HC/FT with participation from Valor Equity Partners, Battery Ventures, Salesforce Ventures, and Hetz Ventures, stacked on top of a $30M Seed, signals long-term intent. This is not fast money chasing surface trends. It is informed capital placing bets on infrastructure that can compound across the startup ecosystem and into the largest enterprises in the world.

Timing is doing heavy lifting here. Enterprises have already spent years centralizing data into warehouses and lakehouses. What they have lacked is a reliable system to turn that stored context into predictive advantage without brittle pipelines or specialized teams for every use case. Fundamental positions NEXUS as that missing layer, a system that meets existing infrastructure where it is and elevates what it can do.

Inside the company, the posture is deliberate. Research depth meets production urgency. Models are not built to impress benchmarks alone but to survive contact with real-world data, messy, incomplete, and high stakes. The team they are assembling reflects that bar. Roles span Forward Deployed Full-Stack Engineers, Forward Deployed Data Scientists, Applied and Research Engineers, ML and Infrastructure Engineers, and Technical Partnerships and go-to-market operators who can translate capability into enterprise impact without losing fidelity.

Fundamental is hiring across these positions for people who want to operate at the point where models influence decisions that move markets. For builders watching the startup ecosystem evolve from generation to prediction, this is one of those moments where the signal is early but unmistakable. Explore roles and see how close you want to get to the future being calculated in real time.