Artificial Intelligence
Data Labeling & RLHF
Data labeling and reinforcement-learning-from-human-feedback (RLHF) infrastructure -- the human-annotation supply chain underpinning frontier AI model training and alignment -- has become a strategically contested, high-margin layer of the AI stack, dominated by Scale AI and Surge AI. Scale AI, now roughly 49%-owned by Meta following a $14.3 billion non-voting-stake investment in 2025 that valued the company at $29 billion, faces internal tension after Meta chief AI scientist Yann LeCun publicly criticized CEO Alexandr Wang's leadership in January 2026; rival Surge AI, bootstrapped and profitable since 2020, hit $1.2 billion in 2024 annualized revenue and began its first external fundraise in July 2025 at reported valuations between $15 billion and $30 billion, positioning itself as the neutral alternative for labs wary of Meta's stake in Scale.
Sources
Every claim traced to a primary source — evidence, recent activity, and insider filing behavior, all in one place.
Evidence
1 source · 46% avg. source confidence
- CNBC / Reuters — Meta-Scale AI deal reshapes data-labeling competitive landscape46% source confidence
Dependency Map
What Data Labeling & RLHF depends on below, and who depends on Data Labeling & RLHF above — click any node to make it the new center, 2 levels deep.
Connection type
Click any node to make it the new center. Scroll to zoom, drag to pan.
The Story So Far
Auto-generated from this entity's dated milestones, relationship updates, and sourced evidence — not AI-written, just sorted.
Data Labeling & RLHF's Timeline
A sourced, dated history of Data Labeling & RLHF's key moments — founding to present.
Jun 2025 · Meta invests $14.3B in Scale AI for 49% non-voting stake
Meta invested $14.3 billion in Scale AI for a roughly 49% non-voting stake, valuing the data-labeling company at $29 billion and installing founder Alexandr Wang in a senior Meta AI role, reshaping competitive dynamics across the RLHF/data-labeling market as rival labs sought alternative vendors perceived as more neutral.