Scale AI
Category: AI Infrastructure
The data infrastructure for AI, providing the high-quality training data and Reinforcement Learning from Human Feedback (RLHF) necessary to power foundation models and autonomous systems. Scale AI was founded in 2016. The company is led by Francis deSouza. Based in San Francisco, USA. Team size: 1,200. Total funding raised: ~$1.6B primary across VC rounds (through Series F); plus Meta $14.3B strategic stake for 49% (Jun 2025, largely secondary liquidity to shareholders) at ~$29B valuation. Latest round: Strategic Investment. Key investors include Accel, Founders Fund, Index Ventures, Tiger Global, Meta, Amazon, Nvidia, Dragoneer.
AMW Analysis
Scale AI provides the data infrastructure layer for AI development, supplying human-verified training data and RLHF services that foundation model and autonomous system developers rely on to reduce hallucination and improve model performance. Founded in 2016 and based in San Francisco, the company has raised roughly $2.6 billion across all rounds, with a $1 billion Series F in April 2026 that valued it at $13.8 billion and drew backing from Amazon, Meta, Nvidia, AMD and Intel Capital to fund a "data foundry" expansion of its labeling pipelines. The most significant development in its recent history is Meta's strategic investment, which acquired a 49% stake for $14.3 billion and pushed valuation estimates as high as $29 billion, though later private-market signals point to a lower figure of roughly $18.9 billion, indicating a gap between strategic and open-market pricing. Its investor base also includes Accel, Founders Fund, Index Ventures, Tiger Global and Dragoneer. Separately, the company has directed capital toward applied AI projects, including a $128.5 million commitment across dozens of initiatives in Canada spanning healthcare, energy and infrastructure applications.
AMW analysis, generated from 4 tracked news signals.
- Founded
- 2016
- Headquarters
- San Francisco, USA
- Team size
- 1,200
- Total funding
- ~$1.6B primary across VC rounds (through Series F); plus Meta $14.3B strategic stake for 49% (Jun 2025, largely secondary liquidity to shareholders) at ~$29B valuation
Value proposition
Accelerates AI development by providing high-quality, human-verified 'ground truth' data at massive scale, significantly reducing model hallucination and improving performance through expert-led RLHF.
Products and solutions
Scale GenAI Platform (RLHF, Red Teaming, and Model Evaluation), Scale Data Engine (High-quality labeling for Vision, NLP, and Audio), Scale Donovan (AI-powered decision-making platform for Federal/Defense), Scale Catalog (Data management and curation for enterprise AI), Scale Forge (Custom data generation and synthetic data services)
Unique value
Acts as the 'foundational layer' of the AI stack by combining a global workforce of subject-matter experts with proprietary ML-assisted labeling tools to create a high-velocity data flywheel.
Target customer
Foundation model labs (OpenAI, Anthropic), autonomous vehicle manufacturers, Fortune 500 enterprises, and government/defense agencies.
Industries served
Generative AI & LLMs, Autonomous Vehicles & Robotics, Government & Defense, E-commerce & Retail, Financial Services, Healthcare
Technology advantage
Proprietary quality-assurance algorithms and a massive scale of human-in-the-loop operations that allow for the rapid fine-tuning of Large Language Models (LLMs) with specialized domain expertise (coding, legal, medical).
How they differentiate
Scale AI differentiates through its 'Data Engine' which integrates a massive human-in-the-loop workforce (Remotasks) with proprietary ML-assisted labeling. It is the market leader in Reinforcement Learning from Human Feedback (RLHF), providing the critical human-verified data required to fine-tune Large Language Models (LLMs).
Main competitors
Labelbox, Snorkel AI, Surge AI, Hive
Key partnerships
Meta (49% non-voting strategic stake; major commercial AI data partner), OpenAI (Data labeling and RLHF), NVIDIA (AI infrastructure), Microsoft, Amazon, Google Cloud (cloud/enterprise distribution and public-sector programs), U.S. Department of Defense / Defense Innovation Unit (Thunderforge and multi-million awards), Qatar government / MCIT (sovereign AI applications partnership)
Notable customers
OpenAI, Meta, Microsoft, Toyota, General Motors, U.S. Department of Defense, Mayo Clinic, BP, Allianz, Cisco
Major milestones
Achieved a $13.8 billion valuation in May 2024 following a $1 billion Series F round., Became a primary data/RLHF partner for major frontier labs including OpenAI., Launched Scale Donovan for federal/defense AI programs., June 2025: Meta $14.3B strategic investment for 49% non-voting stake at ~$29B valuation; founder Alexandr Wang joined Meta; Jason Droege interim CEO., 2025: strongest year — >$1B new business; ~$200M DoD awards including Project Thunderforge; Qatar sovereign AI partnership., March 2026: launched Scale Labs research division expanding SEAL., August 10, 2026: Francis deSouza (ex-Google Cloud COO) appointed CEO.
Growth metrics
2025 strongest financial year: well over $1B in new business (nearly half in Q4); Forbes reports ~$1B revenue in 2025 (up from ~$870M prior year) and expectation to top $1B in 2026. Applications business accelerated; DoD contracts ~$200M in 2025; Meta commercial commitment (~$450M/year for five years).
Market positioning
Dominant category leader in AI data infrastructure and 'Data Foundry' services for foundation model labs and enterprises.
Geographic focus
Headquartered in North America (San Francisco) with a global operational footprint for data labeling across Southeast Asia, Africa, and Europe.
Patents and IP
Holds multiple patents related to automated data annotation, synthetic data generation, and machine learning model evaluation techniques.
About Francis deSouza
Francis deSouza became CEO of Scale AI effective August 10, 2026. Most recently he was Chief Operating Officer and President of Security Products at Google Cloud. Previously President and CEO of Illumina, where he grew revenue to over $4.5 billion and expanded to 150+ countries. Earlier President of Products and Services at Symantec; founded two companies acquired by Microsoft and Symantec. Holds BS and MS in electrical engineering and computer science from MIT.
Latest news about Scale AI
- Scale AI closed a $1 billion Series F round, boosting its valuation to $13.8 billion with new backing from Amazon, Meta, Nvidia, AMD and Intel Capital. The fund will drive its next‑phase "data foundry
- Meta has consolidated its AI future by acquiring a 49 percent stake in Scale AI for 14.3 billion dollars and appointing Dina Powell McCormick to lead capital efforts for infrastructure. 🚀 This integra
- Meta has appointed Scale AI founder Alexandr Wang to lead its new Superintelligence Labs following a landmark 14 billion dollar investment. This move represents a systemic pivot from Yann LeCun academ
- SCALE AI has launched its largest funding round to date, committing 128.5 million dollars to 44 new applied AI projects across Canada. This brings the total investment to 226 million dollars over the
- Meta's $14.3 billion investment in Scale AI in June 2025 catapulted its valuation to $29 billion, a 110% increase from early 2024. However, private market signals show a current valuation around $18.9
More AI Infrastructure companies
Official website: https://scale.com
