What Surge AI does

Surge AI (legal name Surge Labs, Inc.) is a San Francisco-based human-intelligence platform that supplies training, evaluation, and reinforcement-learning data to frontier AI laboratories. Founded in 2020 by Edwin Chen, the company operates a global contractor network of more than 200,000 PhDs across 80+ languages and 500+ disciplines, accessed through a proprietary talent-matching algorithm that assigns contributors based on measured skill rather than credentials. The platform is organized into nine product modules: RL Environments and Agents, Rubrics and Verifiers, RLHF, Supervised Fine-Tuning, Human Evaluation, Expert Professional Domains, Internationalization, Multimodal annotation, and Off-The-Shelf Data. Notable proprietary assets include the EnterpriseBench suite (anchored by CoreCraft, a customer-support simulation with 2,500+ entities), Hemingway-bench (a writing leaderboard), and AdvancedIF (a rubric-based instruction-following benchmark co-developed with Meta).

The technology stack combines the expert workforce with bespoke tooling for rubric design, verifier construction, and RL environment simulation; Surge AI also operates as a research partner, with co-authored publications across instruction finetuning, scalable oversight, constitutional AI, and citation generation with OpenAI, Anthropic, Google, Meta, Stanford, NYU, UC Berkeley, and Princeton. Modalities covered span text, image, video, and audio.

The business model is B2B managed services sold to a small set of named frontier AI labs—OpenAI, Anthropic, Google, Meta, and Microsoft—on usage-based pricing calibrated to expert tier, with rates ranging from $150–$300/hour for university professors to $500–$1,000/hour for Supreme Court lawyers, VC partners, and former startup CEOs. Surge AI was bootstrapped without venture capital, has been profitable from inception, and reported $1.2 billion in 2024 revenue, surpassing Scale AI's reported $870 million. In July 2025 the company entered talks for its first external funding round of up to $1 billion at a $25 billion pre-money valuation with Andreessen Horowitz and Warburg Pincus.

Name

Surge AI

Legal name

Surge Labs, Inc.

Website

https://surgehq.ai

Company type

Private

Founded year

2020

Operating status

Operating

Headcount range

101–250 employees

Short description

Surge AI operates a human-intelligence platform providing RLHF data, expert-curated rubrics, RL environments, and multimodal annotation services to frontier AI labs including OpenAI, Anthropic, Google, Meta, and Microsoft, leveraging a network of 200,000+ PhD contractors.

Ownership category

akta.pro rank

Product category

AI Training Data Platform

NAICS

Computer Systems Design and Related Services (54151), Computer Systems Design and Related Services (5415)

SIC

Services-Computer Processing & Data Preparation (7374), Services-Computer Programming, Data Processing, Etc. (7370)

akta.pro primary industry

AI/ML Engineering & Model Development Services (BPAEAHAF)

akta.pro secondary industries

Privacy-Preserving Data (De-identification, Anonymization, PII Redaction) (HDAAALAG), Model Deployment, Serving & Inference Platforms (HDAAABAF)

Keywords

• AI data labeling
• RLHF data generation
• Human intelligence platform
• AI training data
• Expert annotation services

Headquarters

HQ city

San Francisco

HQ country

United States

HQ region

North America

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Personnel, Technology or R&D, Operations, Marketing or Sales

Revenue model

1. Data Labeling Services: B2B data labeling and human intelligence services for AI model training. Surge AI generated $1.2 billion in revenue in 2024, making it likely the highest-grossing data labeling company globally without any external investment.

Pricing tiers

Go-to-market motion2 records

Distribution channels1 record

Marketing channels4 records

Core offering

Surge AI provides human intelligence services for training and evaluating advanced AI models through a platform that combines RL environments, expert-curated rubrics, RLHF data, SFT demonstrations, human evaluation, and access to a global network of 200,000+ PhD-level experts. The platform supports multimodal annotation across text, image, video, and audio, and serves frontier AI labs with usage-based pricing tied to expert hourly rates ranging from $150 to $1,000 per hour.

Product overview

Surge AI is a data platform providing comprehensive human intelligence services for training and evaluating advanced AI models. The offering consists of nine integrated service modules: RL Environments and Agents (for agentic AI training), Rubrics and Verifiers (for evaluation), RLHF (preference/reward data), SFT (demonstration data), Human Evaluation (expert assessment), Expert Professional Domains (specialist contributors), Internationalization (multi-language data), Multimodal (image/audio/video), and Off-The-Shelf Data (pre-built datasets). These services work together to provide the complete data pipeline needed for frontier AI development, from initial training through evaluation and iteration.

Differentiator

Problem solved

Functional benefit

Products and services

• RL Environments and Agents Rich, complex reinforcement learning environments that challenge agentic AI models in novel ways, featuring expert-designed verifiers that reward AI agent behavior across multi-step tasks. Designed for frontier AI labs training agentic systems.
• RLHF Data Generation Reinforcement Learning from Human Feedback data generation service that embodies the richness and subtleties of human preferences and rewards for training AI models. Used by frontier AI labs to align model behavior with human values.
• Expert Professional Domains Access to elite experts including doctors, lawyers, investment bankers, Fields Medalists, and Harvard professors across STEM and humanities for specialized AI training data. Expert hourly rates range from $150 to $1,000 depending on domain and seniority.
• Internationalization Data Services Data services spanning over 70 languages, teaching AI not just language but cultural values, grammar, idiom, and worldview specific to each locale for global AI deployment.
• Off-The-Shelf Datasets Pre-built datasets spanning RL environments, coding, and core AI capabilities available immediately, including thousands of hours of expert reasoning data for rapid model development.
• EnterpriseBench Suite Agentic RL environment suite including CoreCraft, a fully operational enterprise simulation of a customer support organization with over 2,500 entities and 14 entity types, designed for training and evaluating AI agents in complex business scenarios.
• Hemingway-bench AI writing leaderboard powered by expert human writers evaluating frontier models on real writing tasks, providing gold-standard assessment of AI-generated text quality.

Quantifiable outcome

• $1.2 billion revenue in 2024
• +2 more outcomes

Named customers5 records

Segments1 record

Ideal customer profiles1 record

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability10 records

Feature6 records

Partnerships

Eight partnerships are on record, tiered core.

• MetacoreStrategic or Co-development PartnerResearch partnership on AdvancedIF benchmark for instruction following evaluation and Rubric-Based Benchmarking. Meta Superintelligence Labs partnered with Surge AI to build AdvancedIF using human-crafted rubrics as reward signals for reinforcement learning, achieving 13% gain in instruction-following performance.
• GooglecoreStrategic or Co-development PartnerResearch collaboration on Scaling Instruction Finetuned Language Models. Google partnered with Surge AI to explore instruction finetuning with focus on scaling tasks, model size, and chain-of-thought data.
• OpenAIcoreStrategic or Co-development PartnerMultiple research collaborations including GSM8K dataset of 8,500 math problems and PALMS (Process for Adapting Language Models to Society) with Values-Targeted Datasets.
• AnthropiccoreStrategic or Co-development PartnerCollaborations on Scalable Oversight for large language models, Discovering Language Model Behaviors with Model-Written Evaluations, and Constitutional AI development.
• Stanford UniversitycoreStrategic or Co-development PartnerResearch collaboration on Learning to Reason With Relational Abstractions.
• NYUcoreStrategic or Co-development PartnerJoint research on Inverse Scaling and Learning to Reason With Relational Abstractions.
• UC BerkeleycoreStrategic or Co-development PartnerCollaboration on DOC (Detailed Outline Control) framework for improving long-range plot coherence in AI-generated stories.
• Princeton UniversitycoreStrategic or Co-development PartnerResearch collaboration on enabling large language models to generate text with citations.

Scale indicators9 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• Scale AI: Largest data labeling and RLHF provider for AI labs, with reported 2024 revenue of $870M. Directly competes with Surge AI for the same frontier-lab customers (OpenAI, Anthropic, Meta, Google, Microsoft) using a combination of contractor workforces and AI-assisted labeling tooling.
• Mercor: AI training data and talent platform that connects domain experts (engineers, scientists, lawyers) with AI labs for RLHF and evaluation work. Recently reported at a $10B valuation and is positioned as a direct competitor for the same expert-labeling spend Surge AI captures.
• Labelbox: Data labeling platform offering annotation tooling plus managed labeling services for ML teams. Competes with Surge AI for enterprise AI training budgets, though typically with a software-led rather than expert-network-led value proposition.
• CloudFactory: Managed workforce provider delivering AI data labeling, annotation, and moderation services for ML teams. Comparable in using a distributed contractor model to serve ML training pipelines, though focused on volume labeling rather than Surge AI's expert-tier positioning.
• Sama: AI data annotation provider combining a managed workforce with computer vision and LLM data services. Competes for AI training data budgets with both standard and high-quality labeling tiers, overlapping with Surge AI in serving frontier model developers.

Broad incumbents

• Appen: Public, large-scale data annotation and dataset services company with a global crowd workforce. Overlaps with Surge AI in providing labeled data for ML/AI, but serves a broader enterprise customer base with more commoditized offerings and lower expert-tier specialization.
• TELUS International (formerly Lionbridge AI): Large incumbent digital services provider offering AI data annotation as part of a broader BPO portfolio. Competes for AI training data contracts at scale, though typically with lower-tier, higher-volume labeling rather than Surge AI's expert network focus.

Emerging players

• Snorkel AI: Data-centric AI platform that combines programmatic labeling with weak supervision and expert input. Adjacent competitor to Surge AI in the training-data layer, differentiated by automated labeling approaches versus Surge's expert human network.
• SuperAnnotate: AI data annotation platform providing annotation tooling plus an integrated contributor marketplace. Targets similar ML teams and AI labs for labeled training data, with partial overlap on multimodal and RLHF-adjacent use cases.
• Encord: Data-centric AI platform for annotation, evaluation, and curation of multimodal training data. Competes for labeling and evaluation spend from AI teams building computer vision and multimodal models.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/HelloSurgeAI
• LinkedInlinkedin.com/company/surge-ai
• Facebookfacebook.com/HelloSurgeAI

Revenue estimate

Valuation estimate

Number of profiles

Profiles1 record

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Surge AI