AI labs and model teams
Groups training or evaluating models who can describe the real-world data they need.
For data buyers
Tell us what real-world data your models need
Big Data/Little Guy takes requirements from AI labs, model teams, and other buyers of business or operational data. One anonymized commercial archive is described below. It is not a complete catalog. Availability and rights are reviewed before anything moves forward.
Tell us what data you need
Loading your request.
Who this is for
AI labs, model teams, and other data buyers.
Product and research teams
Teams building agents, tools, or benchmarks who need operational examples rather than a scrape of the public web.
Other dataset buyers
Companies that need business or operational data for their own work, including teams that are not foundation-model labs.
How a request moves forward
Requirements first. Rights before any use.
01
Describe the requirement
Share the categories, domain, and intended use, plus any scope, volume, time, geography, or language constraints you already know.
02
We look for a possible match
A next step depends on availability and a rights review. We do not promise that a matching dataset exists.
03
Samples and identity stay controlled
A controlled sample, or the seller's identity, requires a mutual NDA and the seller's approval when either one is appropriate.
04
A license comes before use
Training or testing waits for a signed license. A request, a conversation, or a profile review does not grant those rights.
One anonymized archive
Commercial drone, geospatial, and operational captures.
The archive figures are approximate and seller-reported. They are not a price, a public sample, or permission to train.
These cards are collections inside one anonymized archive from commercial drone work, not eight sellers and not eight separate datasets. The figures are approximate seller-reported working numbers. Categories overlap, and companion files are counted in more than one group, so the sizes must not be added together.
The seller-reported consolidated working total is about 5 TB, of which the commercial video archive is about 3.3 TB. Video history is reported as 2017–2026. The timeline for the other collections was not established.
The seller reports original commercial capture across roughly nine years of field work in the United States, including construction, energy, industrial, and roofing sites. That is not a claim that every file is cleared, scrubbed, or available to license.
An original representation window ran through about mid-December 2026. That date is not a promise that licensing is available now. A sample or the seller's identity still requires a mutual NDA and the seller's approval. The archive is not described as already scrubbed or pre-cleared. Training or testing waits for a signed license.
Seller-reported thermal imagery is mainly roofs, building facades, and equipment rather than people. That does not mean the files have already been reviewed or scrubbed.
Collection in this archive
Commercial video archive
Long-form produced videos and thousands of short source clips, reported as MP4 and MOV.
- Files
- About 8,900 files
- Size
- About 3.3 TB
- Years
- 2017–2026
- Construction
- Industrial and manufacturing
- Energy and heavy operations
- Commercial buildings and roofing
- FPV interiors
- Real estate and land
- Infrastructure
Collection in this archive
Point cloud and 3D
Seller-reported point-cloud and 3D working files. Format sizes are approximate.
- Files
- 30,523 files
- Size
- About 84.5 GB
- Years
- Years not established
- RCS, about 47.7 GB
- LAS, about 10.7 GB
- OBJ, about 7.6 GB
- PLY, about 6.9 GB
- Also PTX, LAZ, RCP, FBX, and 3D tiles
Collection in this archive
Mapping rasters
TIF and TIFF orthomosaics and other mapping outputs.
- Files
- 677 files
- Size
- About 165 GB
- Years
- Years not established
- 51 mapping-named folders
- 26 ortho-named folders
Collection in this archive
CAD and BIM
Native CAD and BIM files, including survey, terrain, and elevation exports.
- Files
- 378 files
- Size
- About 0.4 GB
- Years
- Years not established
- 325 SKP
- 36 DWG
- 15 DXF
- 2 RVT
Collection in this archive
GIS and geospatial
Geospatial vectors and sidecars from the same archive.
- Files
- 720 files
- Size
- About 0.8 GB
- Years
- Years not established
- KML and KMZ
- SHP
- GeoJSON and GPKG
- TFW, PRJ, and other sidecars
Collection in this archive
Thermal and infrared
15 seller-reported thermal groups. These sit inside the same archive and overlap other groups.
- Files
- File count not established
- Size
- About 141 GB
- Years
- Years not established
- Educational and institutional, about 72 GB across 3 sets
- Compressor stations, about 33 GB across 4 sets
- Industrial and coastal, about 26 GB across 3 sets
- Commercial buildings and roofing, about 9 GB across 3 sets
- Solar, energy, and other, about 1 GB
Collection in this archive
Specialty sets
Named working sets inside the archive. Their sizes are already part of the overlap.
- Files
- File count not established
- Size
- About 456 GB
- Years
- Years not established
- Multi-year construction progress, about 307 GB
- Offshore energy operations, about 79 GB
- Oil and gas and energy, about 52 GB
- Multi-date construction and mapping, about 11 GB
- Commercial construction documents, about 5 GB
- Roofing and thermal, about 2 GB
Collection in this archive
Mapping groups
12 seller-reported mapping groups inside the same archive.
- Files
- File count not established
- Size
- About 167 GB
- Years
- Years not established
- Engineering and land mapping, about 141 GB
- Environmental and wetlands, about 6 GB
- Multi-date construction, about 6 GB
- Large-acreage orthomosaics, industrial line-of-sight, corridor survey, and others, about 14 GB
What these cards do not grant
- A complete catalog, a price, or a promised timeline
- Files that are already scrubbed or pre-cleared for every use
- Permission to train, test, or receive raw files from this page