Kled is a research consortium of top AI scientists focused on forecasting AI data scarcity and, through our mobile application layer, collecting the high-value human datasets required to push the frontier forward.

We are backed by early investors behind SpaceX, Airbnb, Coinbase, xAI, OpenAI, Anthropic, Higgsfield AI, Spotify, Lyft, Uber, and more.

To date, Kled has built the largest explicit opt-in network of uploading data contributors on the planet. Charting number 1 on the App Store for finance in 4 countries and garnering hundreds of thousands of users in 6 weeks.

Thesis

1. Data is the bottleneck.

   Model progress is no longer constrained by architecture, but by access to high-quality, human-generated data. Something as simple as camera roll data can have a meaningful impact on visual model realism - Our Full Research Paper

   
   

2. The internet is exhausted.

   Scraped data is finite, low-signal, and increasingly synthetic. The next gains require fresh, opt-in human data.

   
   

3. Rights-aware data wins.

   The future of AI training is compliant, consent-based, and economically aligned with the people who create the data. Lawsuits have forced the proper licensing of data for training.

   
   

4. Consumers are the supply side.

   The highest value datasets emerge from everyday human behavior, not enterprise silos.

   
   

5. Scale requires incentives.

   You don’t collect frontier data by asking nicely. You build systems where contributors are paid continuously.

Network

Consumer Network

We have built the largest opt-in data collection effort since the birth of the internet. Our app lets anyone upload their personal data (images, videos, text documents, etc) and get paid for it.

• ≈200,000 daily active users hit in 30 days of launching the app. Avg 3-4.5 million files are uploaded and processed every day. 100x-200x more uploads per day per user compared to both Facebook and Instagram combined - calc.

  
  

• 150,000,000 impressions on social media across all platforms.

  
  

• 3x daily active users day over day, 2x upload rates day over day during our initial launch period. Once a user has been paid twice for uploading their data, we have them for life.

  
  

• Number 1 on the App Store with 0 marketing (Finance Category), charting in several countries

Enterprise Network

Beyond our consumer application, Kled has partnered with hundreds of leading rightsholders across film, production, healthcare, finance, engineering, etc. This is low-hanging fruit for us because of our strong rights holder relationships. Full Dataset Breakdown

We’ve signed exclusive deals to own the exclusive AI licensing rights for:

• 140,000+ mainstream IPs, including episodic film and television content from Bugs Bunny, Tom and Jerry, Garfield, Superman, The Twilight Zone, Mighty Mouse, and other top-tier catalogs

• 50,000,000 anonymized patient medical records, sourced from large U.S. healthcare and medical billing infrastructure providers

• POV military and real-world warfare footage, captured from first-person perspectives

• Egocentric POV datasets for robotics training, including cooking, household chores, manufacturing workflows, and real-world task execution

• Enterprise internal communications datasets, including over anonymized Slack conversations from engineering, consulting, accounting, and professional services firms

• Enterprise operational datasets, spanning SAP ERP systems, Oracle NetSuite, Shopify commerce data, and other large-scale business infrastructure

Product

Kled is the filtering layer. Our job is to take low-signal, unconstrained human participation and convert it into high-utility, rights-cleared datasets that resolve real data scarcity at the frontier.

Kled does this by mobilizing its consumer network as an adaptive supply layer. Rather than passively accepting uploads, we actively steer participation toward data that is scarce, underrepresented, or newly valuable to modern models. This is accomplished through special tasks, incentive shaping, and real-time demand signals from enterprise and research buyers.

Turning Data High Signal (Post Upload)

Data that is contributed is automatically subjected to a layered filtering and enrichment pipeline:

All duplicate uploaded content is rejected, all internet-plagiarized content is rejected, all synthetic data is rejected, all misclassified data is rejected, all altered metadata is rejected.

Data that gets accepted through this second filtration layer is turned into standardized, labeled, model-ready data that can be deployed immediately for training, evaluation, or fine-tuning.

Customers

Kled is enriching and supplying data for top enterprises across healthcare, shopping, transportation, advertising and more. Superpower, Nexar, Dupe, etc.

The growth of our newly launched consumer app has opened doors into the world’s leading AI labs and data resellers. We're backed by the founder of Waymo, the head of drones at DoorDash, the founder of LVMH's venture fund, and investors behind xAI, OpenAI, Anthropic, Higgsfield AI, etc.

We've paid out millions of $ to users and have $30 million + in the pipeline to purchase the data our users are uploading.

Team

Team composed of Research, Product, & Growth teams. Headquarters moving from NYC to San Francisco, in-person & remote teams, work 9-9/7:

Applied AI & Research: Ex-NASA and Stanford University applied AI research graduates. Built vision training pipelines spanning egocentric video, long-tail image distributions, and real-world sensor data; worked on diffusion and adapter-based fine-tuning, dataset influence scoring, and empirical evaluation of data quality vs. model capability.

Data Systems & Infrastructure: 4× technical founder, built over 200 mobile and web applications with millions of users. Highest-earning Framer developer (2024–2025). Ex-Amazon founding engineer team, scaled systems to handle petabytes of data.

Growth & Sales: 3x exited founders, built IP and licensing marketplaces to 400M+ assets and $1B+ in rights, and creators with 2M+ subscribers and 10B+ views across platforms.

Legal Representation - Latham & Watkins + Mission Law + Fenwick

2026 (What's Next)

Our goal is to create the first $10,000 per month data uploader by end of Q1. Our top user is making ≈$7,400 per month. With our average top 5% user making ≈$600 per month.

Creating this user will be the result of more optionality via the app. More features + Tasks for our users.

In addition to this optionality, we need to tighten up our core fraud detection systems; complex tasks will require better fraud detection systems to detect poor quality uploads for hyper-specific specs/tasks.

Lastly, we're focused on our Android app release, which will comfortably see 100x more growth than our iOS app.

Latest Raise

Seed Round: $5.5mil+ raised from K5 Global, Parable VC, Sebastian Thrun (Founder of Waymo), Cox Exponential (CX2), Aglaé Ventures (Venture Fund of Bernard Arnault), Slow Rush Ventures (Abe Burns), Luca Netz (Founder of Pudgy Penguins), Daniel Greenberg (Founder of Mschf), Harrison Shih (Head of Drones at DoorDash), Diplo, 24kGoldn, Kent Lacob (Owner of the Golden State Warriors) and many other notable angles.

Other Investors: Zach Greenberger (CBO of Lyft), Max Marchione (Founder of Superpower), Paul Yacoubian (Founder of Copy AI), Everbloom, Sean Li (Founder of Magic Labs), A16Z Scout, Bain Scout, creators with 100+ million followers like Panda Boi, and more.