Poseidon
Poseidon is a full-stack AI data infrastructure company that sources and licenses rights-cleared training data, primarily audio, to AI labs and model developers through its B2B platform and consumer Numo contributor app.
- Company typePrivate
- Founded2025
- HeadquartersSan Francisco, United States
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What Poseidon does
Poseidon AI, Inc. is a San Francisco-based, venture-backed data infrastructure company founded in 2025 that builds a full-stack decentralized layer for sourcing and licensing rights-cleared training data for AI systems. The company operates two interconnected products: a B2B data licensing platform that sells structured, commercially licensed datasets to AI labs and model developers, and a consumer mobile application called Numo that crowdsources data contributions from individual users in exchange for blockchain-recorded rewards. Data is collected with contributor consent, curated with statistical outlier flagging and deepfake quality scoring, labeled using a mix of AI and human consensus annotations, and licensed on-chain via Story Protocol's infrastructure, which provides provenance and audit trails that protect developers from intellectual property liability.
The company's core technical differentiator is its focus on long-tail, non-European languages where it claims to exceed major public datasets by an order of magnitude, covering Indonesian, Vietnamese, Spanish, Hindi, Russian, Korean, Urdu, Portuguese, Mandarin, French, German, Arabic, Turkish, Japanese, and Marathi. As of the input it has compiled 2,500,000+ audio files spanning 16,500+ hours across 8+ languages. Poseidon was incubated by the DATA Foundation (formerly Story Protocol / PIP Labs) and raised a $15 million seed round led by Andreessen Horowitz's crypto fund in July 2025. In May 2026 it launched Numo, which reached 15,000+ contributors and 170,000+ recordings in its first week, and in June 2026 it partnered with Toss, South Korea's national mobile financial platform, embedding Numo as a mini-program inside Toss's app to access roughly 30 million users for Korean voice, image, and behavioral data collection.
Commercially, Poseidon licenses rights-cleared training datasets to AI labs working on speech transcription, humanoid robotics, autonomous vehicles, and multi-modal pre-training; pricing is not publicly disclosed and appears quote-based. Revenue mechanics combine data licensing (license/royalty-style) with transaction fees flowing through the Numo contributor platform, though current revenue is undisclosed given the company is in its first year of go-to-market. The company is privately held with offices in San Francisco, Palo Alto, and Seoul.
Poseidon firmographics
Firmographics- Name
- Poseidon
- Legal name
- Poseidon AI, Inc.
- Website
- https://psdn.ai
- Company type
- Private
- Founded year
- 2025
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- Poseidon is a full-stack AI data infrastructure company that sources and licenses rights-cleared training data, primarily audio, to AI labs and model developers through its B2B platform and consumer Numo contributor app.
- Ownership category
- akta.pro rank
Poseidon industry classification
Industry- Product category
- AI Training Data Infrastructure
- NAICS
- Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518)
- akta.pro primary industry
- Ethical Data Sourcing, Consent & Data Provenance (HDAAAMAI)
- akta.pro secondary industries
- AI Privacy Engineering & Data Governance (PII, consent, retention) (HDAAAKAB), Data Privacy, Consent & Compliance Management (HDAEADAG), Privacy-Preserving Data (De-identification, Anonymization, PII Redaction) (HDAAALAG), Audit, Explainability & Accountability Tooling (traceability, reporting) (HDAAAKAL)
Keywords
Where Poseidon is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Offices3 records
Markets served
Poseidon business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Infrastructure, Operations, Marketing or Sales
Revenue model
- Data Licensing: Poseidon licenses structured datasets collected with consent, curated for quality, and licensed for commercial use to AI labs and model developers. The company earns revenue by selling access to high-quality, rights-cleared training data.
- Contributor Platform Fees: Poseidon operates the Numo platform where users contribute data and receive blockchain-recorded rewards. The platform facilitates real-time data payouts for contributors.
Go-to-market motion2 records
Distribution channels2 records
Marketing channels4 records
Poseidon product offering
Product offeringCore offering
Poseidon is a full-stack data layer for AI development that sources, curates, labels, and licenses rights-cleared training datasets for AI labs and model developers. The company operates the Numo consumer mobile app to crowdsource real-world audio and behavioral data, then processes it through AI-assisted curation and consensus human labeling before licensing the resulting datasets commercially. Data provenance and rights clearance are recorded on-chain via Story Protocol, enabling enterprise AI buyers to train and fine-tune models without IP liability.
Product overview
Poseidon operates as a full-stack AI data infrastructure company offering two interconnected products: (1) Numo, a consumer application for crowdsourced data contribution with real-time blockchain-based rewards, and (2) a B2B data licensing platform providing rights-cleared, commercially licensed training datasets for AI developers. The platform bridges supply and demand for long-tail, rights-cleared training data through a complete pipeline of collection via Numo contributors, quality curation with AI-assisted and human annotation, and on-chain licensing via Story Protocol for data provenance and rights management. Poseidon targets AI labs requiring high-quality, IP-safe training data and differentiates through its scale (2.5M+ audio files, 16,500+ hours across 8+ languages), quality verification with deepfake detection scoring, and blockchain-based rights management.
Differentiator
Problem solved
Functional benefit
Brands
- Numo: Consumer application developed by Poseidon that enables users to contribute real-world, rights-cleared training data for AI systems through mobile tasks, starting with voice recordings in underrepresented languages.
Products and services
- Numo Consumer mobile application developed by Poseidon that enables individual users to contribute real-world voice and behavioral data through mobile tasks in exchange for blockchain-recorded rewards. The app launched as a mini-program within the Toss app, giving it access to 30 million South Korean users for crowdsourced Korean-language data collection, with broader multilingual expansion planned.
- Data Licensing Infrastructure B2B platform providing rights-cleared, commercially licensed training datasets for AI developers, integrating Story Protocol's on-chain licensing for data provenance and audit trails. The platform supports AI workflows for speech transcription, humanoid robotics, autonomous vehicles, and multi-modal pre-training, with 2,500,000+ audio files and 16,500+ hours across 8+ languages, including a depth advantage in low-resource languages.
Quantifiable outcome
- 15,000+ contributors and 170,000+ voice recordings in under a week of Numo launch
Companies that use Poseidon
Customer profileSegments1 record
Ideal customer profiles1 record
Poseidon technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- No
- API docs
- API detail
Core technology
AI maturity
App detail
AI capability5 records
Feature3 records
Poseidon partnerships and signals
Strategic signalPartnerships
Two partnerships are on record, tiered flagship and core.
- Toss (Viva Republica)flagshipPoseidon partnered with Toss, South Korea's national mobile financial platform, to integrate its Numo contributor app as a mini-program within the Toss app. This partnership gives Poseidon direct access to Toss's approximately 30 million users, enabling them to generate Korean voice, image, and POV behavioral data within the app while receiving real-time blockchain-recorded rewards. The collaboration aims to prove the data sourcing model in Korea before expanding globally.
- DATA Foundation (formerly Story Protocol)corePoseidon is a core ecosystem AI project of DATA Foundation (formerly Story Protocol). The company was incubated by the DATA Foundation team and integrates Story Protocol's onchain licensing infrastructure for data provenance, licensing, and auditability. Poseidon leverages this infrastructure for transparent, blockchain-verified rights management and contributor compensation.
Scale indicators7 records
Recent moves5 records
Expansion highlights6 records
Poseidon competitors and assessment
Company assessmentDirect peers
- Scale AI: Scale AI is the dominant provider of training data for AI labs, offering data annotation, collection, and evaluation services across multiple modalities. Directly comparable to Poseidon's B2B data licensing model for AI developers, though Scale focuses more on annotation services and synthetic data while Poseidon emphasizes rights-cleared provenance.
- Defined.ai: Defined.ai (formerly DefinedCrowd) sells pre-collected, ethically-sourced training datasets for speech, NLP, and computer vision to AI developers. It is a direct competitor to Poseidon's model of licensing ready-made, rights-cleared datasets rather than offering annotation services.
- Surge AI: Surge AI provides high-quality human-labeled training data for frontier AI labs, particularly for RLHF and nuanced annotation tasks. It is comparable to Poseidon as a premium data provider targeting the same AI lab buyer segment, though Surge focuses on labeling quality while Poseidon focuses on rights-cleared sourcing.
- CloudFactory: CloudFactory provides AI data annotation services through a managed workforce, including audio transcription and image labeling. It is comparable to Poseidon as a supplier of human-in-the-loop training data, though CloudFactory emphasizes workforce management over rights-cleared provenance.
- Sama: Sama provides AI data annotation and validation services with a focus on ethical sourcing and impact employment. It is comparable to Poseidon as a rights-conscious training data supplier serving AI labs, though Sama emphasizes annotation workflows while Poseidon focuses on crowdsourced data collection and licensing.
Broad incumbents
- Appen: Appen is a publicly-traded incumbent in the AI training data market, offering annotated datasets across 235+ languages with a global contributor crowd. It competes with Poseidon for the same multilingual, rights-cleared training data opportunity but operates at vastly greater scale and with established Fortune 500 customer relationships.
Emerging players
- Toloka: Toloka crowdsources data labeling and collection through a global network of contributors, with particular strength in low-resource and non-English languages. It overlaps with Poseidon's focus on multilingual training data sourced from a distributed contributor base, though Toloka's model is more labeling-centric.
- Snorkel AI: Snorkel AI offers programmatic data labeling and weak supervision approaches to training data preparation for AI developers. It competes with Poseidon in serving AI labs' data needs, though Snorkel's approach centers on synthetic labels from subject-matter experts rather than crowdsourced raw data collection.
- Labelbox: Labelbox provides a data-centric AI platform for annotation, curation, and evaluation of training data, serving AI developers building production models. It competes with Poseidon for AI lab data infrastructure spend, though Labelbox focuses on tooling rather than owning/licensing the underlying data.
Others
- Common Voice (Mozilla Foundation): Common Voice is Mozilla's open-source, crowdsourced multilingual speech dataset that Poseidon explicitly benchmarks against. While not a commercial competitor, it represents the free public alternative that constrains pricing power and forces Poseidon to justify its premium via rights clearance and quality.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat5 records
Key risks6 records
Key highlights7 records
Customer concentration
Poseidon social profiles
Digital presencePoseidon financial estimates
Financial estimateRevenue estimate
Valuation estimate
Poseidon leadership team
Management profileNumber of profiles
Profiles5 records
Poseidon funding detail
Funding detailFunding overview
Funding rounds1 record
Investors1 record
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Poseidon M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Poseidon
What does Poseidon do?
Poseidon is a full-stack data layer for AI development that sources, curates, labels, and licenses rights-cleared training datasets for AI labs and model developers. The company operates the Numo consumer mobile app to crowdsource real-world audio and behavioral data, then processes it through AI-assisted curation and consensus human labeling before licensing the resulting datasets commercially. Data provenance and rights clearance are recorded on-chain via Story Protocol, enabling enterprise AI buyers to train and fine-tune models without IP liability.
Is Poseidon a public or private company?
Poseidon is a private company. It is classified as venture growth investor backed and is currently operating.
When was Poseidon founded?
Poseidon was founded in 2025. It employs 11 to 50 people.
Where is Poseidon based?
Poseidon is headquartered in San Francisco, United States, in the North America region.
How does Poseidon make money?
Two revenue lines are on record. Data Licensing is the primary driver. The others are contributor Platform Fees.
Who are Poseidon's main competitors?
Direct peers on record are Scale AI, Defined.ai, Surge AI, CloudFactory and Sama. Appen is listed as a broad incumbent. Emerging players are Toloka, Snorkel AI and Labelbox. Common Voice (Mozilla Foundation) is listed as an others.
Does Poseidon have an API?
No public API is recorded for Poseidon.
What industry is Poseidon in?
Poseidon's product category is AI Training Data Infrastructure. Its primary akta.pro industry code is HDAAAMAI, Ethical Data Sourcing, Consent & Data Provenance, with a secondary code of HDAAAKAB, AI Privacy Engineering & Data Governance (PII, consent, retention). Its NAICS code is 518.