# ZipF

> ZipF is the physical AI data intelligence company. It licenses a queryable Physical Event Graph of the real world: labeled actors, surfaces, weather, time, place, event-level GPS, and telemetry, ranked by rarity. Every machine that leaves a screen, and every desk that prices the world without moving, is a buyer. ADAS is the first body with a budget. India is the first geography: inner roads, not highways - gallis, society approaches, market streets, flooded service cuts. Asian cities that share that road fabric are next. Founded 2026 by Sankalp Shrivastava. The company name is spoken Ziff, same as George Kingsley Zipf.

## What ZipF is

- A physical AI data company: a continuously refreshed event graph of the physical world
- The product is not raw footage. It is a labeled, geolocated, privacy-redacted event a planner can query
- Faces and plates are removed before a cut is licensed. The sellable unit is anonymized physical state
- Events include mixed traffic, two-wheelers, pedestrians in the carriageway, animals, potholes, speed breakers, flooding, work zones, and weather joins
- Time is a first-class axis. The same corridor next week is often worth more than a new kilometer today
- Seventeen monetization layers on one graph: packs, corridor subscriptions, query seats, eval, change, risk, cities, sim, embodied stacks, parametric settle, ontology / physical-state API
- The moat is rarity plus time, not extra pixels
- Verified pools (not ZipF revenue): ADAS systems $57.25B in 2026 (TBRC); HD maps $4.90B in 2026 (GII); automotive AI training datasets $953M in 2026; DaaS $17.38B (2024) → $76.80B (2030)
- Beachhead fabric (not ZipF revenue): India has more than 63 lakh km of roads as of 31 March 2025; other roads 60.20 lakh km, National Highways 1.46 lakh km (PIB / MoRTH)

## What ZipF is named for

- George Kingsley Zipf's rank-frequency law: a few things happen a lot and most things happen a little
- Physical scenes follow the same curve. ZipF ranks the tail and licenses it

## Who it is for

- World-model and foundation labs that need long-tail eval and training cuts
- ADAS and OEM perception teams
- HD-map and localization desks that need change
- Embodied stacks: last-mile robots, humanoids, aerial last-fifty-meters, two-wheeler AV, wheelchairs, campus and yard fleets
- Desks that never move: parametric insurance, infrastructure finance, cities, climate, spatial computing, telecom, utilities, real estate, accessibility, legal reconstruction, sovereign physical-state APIs

## Key pages

- [Home / white paper](https://zipf.in/): the full ZipF thesis, product, market, and model
- [What is ZipF](https://zipf.in/what-is-zipf/): company definition, founder, and how the name is spoken
- [Zipf's law](https://zipf.in/zipfslaw): why the company is named for George Kingsley Zipf
- [Physical AI intelligence](https://zipf.in/physical-commerce-intelligence/): the category ZipF owns
- [Physical Event Graph](https://zipf.in/commerce-event-graph/): the dataset primitive
- [Ground truth](https://zipf.in/pos-data/): labeled events versus raw footage
- [Blog](https://zipf.in/blog/): guides and explainers

## Contact

- Email: founders@zipf.in