Home  >  Companies  >  Surge AI
Surge AI
Data-labeling platform for AI developers to train models with high-quality annotated datasets

Revenue

$1.20B

2024

Funding

$1.00B

2025

Details
Headquarters
San Francisco, CA
CEO
Edwin Chen
Website
Milestones
FOUNDING YEAR
2021
Listed In

Revenue

Sacra estimates that Surge AI generated $1.2 billion in revenue in 2024, with growth mainly driven by a group of ~12 frontier AI labs, most notably OpenAI, Google, Anthropic, Microsoft, and Meta, as demand has grown for expert data labeling for training next-generation LLMs.

Surge reportedly operates with high capital efficiency, maintaining profitability while managing a global workforce of approximately 50,000 expert contractors and 130 full-time employees. The business model combines usage-based pricing for API access with managed service contracts for complex RLHF projects.

Valuation

Surge AI remains bootstrapped, with no external funding raised since its 2021 launch. As of July 2025, the company was said to be considering its first institutional funding round, targeting $1 billion in new capital at a potential $15 billion valuation.

Product

Surge AI is a data labeling platform that connects AI developers with a curated network of expert human annotators, referred to as Surgers. The platform converts raw text, code, images, and conversation transcripts into structured training data used to train AI models.

Machine learning engineers can create projects through Surge's web interface or Python SDK, designing custom annotation tasks with drag-and-drop tools. They can define specific skill requirements, such as native Spanish speakers with accounting expertise, and upload data that is immediately distributed to matched Surgers globally.

The platform includes tools for reinforcement learning from human feedback (RLHF), enabling live chat evaluation and asynchronous transcript rating. These tools allow humans to rate model responses and provide improved alternatives. Red-teaming workflows support labs like Anthropic in identifying model safety gaps through adversarial testing.

Real-time dashboards monitor annotation quality using metrics such as gold-standard accuracy, inter-annotator agreement scores, and per-worker trust ratings. Labels identified as low quality are automatically reassigned to other annotators, ensuring data integrity at scale.

Surge's workforce is subject to rigorous vetting, including domain-specific tests, background checks, and ongoing performance evaluations. Approved Surgers earn premium rates of 30-40 cents per working minute, exceeding typical crowdsourcing platform rates and incentivizing consistent quality.

Business Model

Surge operates a B2B managed marketplace that generates revenue through usage-based pricing and project-based contracts. Customers are charged per annotation task or can purchase bundles for large-scale RLHF initiatives, with pricing adjusted based on task complexity and the level of domain expertise required.

The platform integrates software-as-a-service tools with a curated labor marketplace, deriving value from both technology licensing and workforce coordination. Enterprise customers have the option to integrate their own data vendor licenses to lower costs while utilizing Surge's orchestration layer.

Surge achieves strong gross margins by managing contractor payments, cloud infrastructure expenses, and platform development costs. Its asset-light model scales efficiently, as the majority of operational costs are variable and directly linked to task volume and complexity.

The business benefits from network effects as the platform attracts more expert annotators and customers increase their usage across diverse AI training workflows. Long-term contracts with frontier labs provide revenue stability, while usage-based pricing captures additional value from expanding model development cycles.

Surge's emphasis on quality enables it to charge premium prices compared to commodity annotation services. Its specialization in domain experts across fields such as mathematics, law, and coding creates switching costs for customers who integrate these workflows into their model training pipelines.

Competition

Vertically integrated platforms

Scale AI is the primary competitive threat, offering end-to-end data infrastructure that integrates compute, evaluation APIs, and human labeling services. Scale uses bundled offerings and subsidized RLHF pricing to maintain market share. However, recent customer defections to Surge suggest concerns about trust and quality.

Meta's 49% stake in Scale, combined with founder departures, has created challenges for Scale and opportunities for Surge to attract customers seeking independent providers.

AWS SageMaker Ground Truth and Google Cloud's Vertex AI now incorporate RLHF templates directly into model training workflows. This integration poses a threat to pure-play providers by enabling self-service labeling within customers' existing cloud environments.

Specialist marketplaces

Micro1 competes by combining AI-powered expert vetting with workforce management, targeting similar high-quality outcomes with faster hiring cycles. Its broader focus on talent automation beyond data labeling could lead to expansion into Surge's core markets.

Prolific operates a 35,000-member vetted panel with granular demographic filtering, offering cost advantages for preference data collection. However, it lacks the domain-specific expertise depth that Surge provides.

Toloka offers global multilingual crowdsourcing with pay-as-you-go pricing, which can undercut Surge on commodity tasks. Despite this, its quality controls and expert curation fall short of Surge's premium service standards.

Self-serve platforms

Labelbox, SuperAnnotate, and Label Studio provide seat-based RLHF pipelines, enabling AI labs to manage expert recruitment internally. These platforms offer cost-effective solutions for organizations with strong internal operational capabilities.

The self-serve model appeals to labs prioritizing control over annotation workflows and data security. However, achieving the same quality and speed as Surge's managed services requires significant internal resources.

TAM Expansion

New products

Surge can expand into AI evaluation and red-teaming suites as continuous model testing becomes mandatory under emerging AI safety regulations. Offering these capabilities as self-serve SaaS tools would enable recurring revenue streams from ongoing model monitoring, supplementing one-time training data creation.

Synthetic data generation presents an opportunity as demand for privacy-preserving training corpora increases. Surge could apply its annotation expertise to audit and improve synthetic datasets, capturing budgets currently allocated to specialized vendors.

The company could also develop reward model APIs and premium preference datasets to complement open-source alternatives, monetizing quality differentiation as labs seek proprietary training advantages.

Customer base expansion

Enterprise adoption of AI in regulated industries is driving demand for compliant annotation services in sectors such as finance, healthcare, and defense. Surge's SOC 2 compliance and security capabilities make it a viable provider for these high-value verticals.

Government AI safety institutes require independent model evaluation services, creating opportunities to compete with Scale in the public sector. Surge's positioning as a neutral, non-conflicted provider could address this need.

Expanding into European markets aligns with AI Act compliance requirements, particularly for human oversight and multilingual capabilities, which Surge's global workforce is equipped to deliver.

Geographic expansion

London's role as Europe's largest generative AI hub, combined with strict data sovereignty requirements, creates demand for regional annotation centers. Establishing EU operations would allow Surge to secure contracts requiring on-shore data handling.

Asia-Pacific markets are experiencing rapid LLM adoption alongside stringent data localization laws. Surge's expert-curation model could meet regional compliance needs while maintaining quality standards.

Partnerships with GPU cloud providers such as CoreWeave could enable bundled dataset and training packages, offering a counter to Scale's compute-subsidized competitive strategy.

Risks

Customer concentration: Surge's reliance on 12 customers for over $1 billion in revenue creates exposure to contract losses and pricing pressure. The departure of a single major customer could materially affect financial performance, while the limited customer base constrains diversification, reducing revenue stability.

Synthetic data substitution: Advances in AI-driven automated labeling and synthetic data generation pose risks to Surge's human-centric business model. As models improve in self-optimization and synthetic training data achieves higher quality, demand for premium human annotation services may decline, potentially undermining Surge's core offering.

Margin compression: Competition from cloud providers bundling annotation tools and self-serve platforms offering lower-cost alternatives could pressure Surge to lower prices to retain customers. The company's high-margin strategy, reliant on quality differentiation, faces challenges as competitors enhance their offerings and customers contend with budget constraints.

News

DISCLAIMERS

This report is for information purposes only and is not to be used or considered as an offer or the solicitation of an offer to sell or to buy or subscribe for securities or other financial instruments. Nothing in this report constitutes investment, legal, accounting or tax advice or a representation that any investment or strategy is suitable or appropriate to your individual circumstances or otherwise constitutes a personal trade recommendation to you.

This research report has been prepared solely by Sacra and should not be considered a product of any person or entity that makes such report available, if any.

Information and opinions presented in the sections of the report were obtained or derived from sources Sacra believes are reliable, but Sacra makes no representation as to their accuracy or completeness. Past performance should not be taken as an indication or guarantee of future performance, and no representation or warranty, express or implied, is made regarding future performance. Information, opinions and estimates contained in this report reflect a determination at its original date of publication by Sacra and are subject to change without notice.

Sacra accepts no liability for loss arising from the use of the material presented in this report, except that this exclusion of liability does not apply to the extent that liability arises under specific statutes or regulations applicable to Sacra. Sacra may have issued, and may in the future issue, other reports that are inconsistent with, and reach different conclusions from, the information presented in this report. Those reports reflect different assumptions, views and analytical methods of the analysts who prepared them and Sacra is under no obligation to ensure that such other reports are brought to the attention of any recipient of this report.

All rights reserved. All material presented in this report, unless specifically indicated otherwise is under copyright to Sacra. Sacra reserves any and all intellectual property rights in the report. All trademarks, service marks and logos used in this report are trademarks or service marks or registered trademarks or service marks of Sacra. Any modification, copying, displaying, distributing, transmitting, publishing, licensing, creating derivative works from, or selling any report is strictly prohibited. None of the material, nor its content, nor any copy of it, may be altered in any way, transmitted to, copied or distributed to any other party, without the prior express written permission of Sacra. Any unauthorized duplication, redistribution or disclosure of this report will result in prosecution.