Nettle
Enterprise data an AI can actually use
PITCHING AT THE SAN FRANCISCO NIGHT ON 25 NOVEMBER 2026
Enterprises and the consultancies that serve them spend most of their engineering time turning fragmented, poorly documented data into something an analysis or an AI can run on. Nettle is an autonomous data platform that ingests raw enterprise data, works out its structure and relationships, flags errors with the evidence, and serves prepared data to dashboards, models and agents.
Every enterprise AI deployment runs into the same wall: the data. Pipelines were built and bolted on one use case at a time, documentation lags reality, and batch jobs that fail halfway leave partial corruption that takes days to unpick. The standard answer is a forward-deployed engineering team that rebuilds the plumbing by hand for each customer, which is slow, expensive and the reason most companies never get past a pilot.
Nettle automates that work. Point it at raw data and it classifies tables, extracts entities and relationships across rows and columns, decomposes dependencies, and detects errors with the evidence for each one, then serves the prepared result through APIs so dashboards, machine learning models and agents query structured data instead of reading raw files. Founder Divye Kapoor leads a team drawn from Pinterest, Google, LinkedIn, HashiCorp and Amazon, alongside a mathematical modelling group of physicists and control theorists. The initial product is live and the company is talking to its first design partners. Divye has pitched at Startup Valley San Francisco twice already and returns in November 2026.
If Nettle is your kind of company.
These founders pitched at the same startup events. The room is usually the reason people find each other.