Prior Labs
Prior Labs builds TabPFN, a family of tabular foundation models that deliver state-of-the-art predictions on structured data in seconds without training or tuning, serving enterprise customers across finance, healthcare, industrials, and technology via API, cloud, and SAP distribution channels.
- Company typePrivate
- Founded2024
- HeadquartersBerlin, Germany
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What Prior Labs does
Prior Labs GmbH is a Freiburg, Germany-based AI research and product company founded in 2024 that builds TabPFN, a family of tabular foundation models purpose-built for prediction on structured data (tables, relational datasets, time series). Its core technology is a transformer architecture pre-trained on millions of synthetic tabular datasets that performs in-context learning — producing state-of-the-art classification, regression, forecasting, and anomaly-detection predictions in seconds without task-specific training, hyperparameter tuning, or feature engineering. The model family has progressed through TabPFN-v2 (Nature publication, January 2025), TabPFN-2.5 (November 2025, 50K rows and 2,000 features), and TabPFN-3 (May 2026, 1M rows natively, plus Thinking mode for test-time compute scaling). A specialized TabPFN-TS variant handles time-series forecasting.
Prior Labs monetizes via multiple deployment surfaces: a hosted Managed Inference API (REST + Python SDK), an enterprise VPC/on-prem offering (TabPFN-3-Plus with Thinking mode) distributed through AWS SageMaker, Azure AI Foundry, Google Vertex AI, Databricks, and Snowflake, and a Model Context Protocol (MCP) server that exposes TabPFN as a callable tool for LLM-based agents. Go-to-market combines a product-led free playground and open-source distribution (3M+ PyPI downloads, 7K+ GitHub stars) with direct enterprise sales. Named customers span finance (TD Bank, Creditplus Bank, Taktile), healthcare (BostonGene, Oxford Cancer Analytics, University Hospitals of North Midlands NHS Trust), industrials (Hitachi Rail), tech (Databricks, Exito), and government (European Commission).
On May 4, 2026, Prior Labs entered a definitive agreement to be acquired by SAP SE, with SAP committing more than €1 billion over four years to scale Prior Labs into a globally leading frontier AI lab for structured business data. Prior Labs will continue to operate as an independent legal entity within SAP, retaining brand, Freiburg headquarters, research mission, open-source strategy, and existing customers, while integrating with SAP Business Data Cloud and the Joule agent stack. The transaction is pending regulatory approval and is expected to close in Q2 or Q3 2026.
Prior Labs firmographics
Firmographics- Name
- Prior Labs
- Legal name
- Prior Labs GmbH
- Website
- https://priorlabs.ai/
- Company type
- Private
- Founded year
- 2024
- Operating status
- Acquired
- Headcount range
- 11–50 employees
- Short description
- Prior Labs builds TabPFN, a family of tabular foundation models that deliver state-of-the-art predictions on structured data in seconds without training or tuning, serving enterprise customers across finance, healthcare, industrials, and technology via API, cloud, and SAP distribution channels.
- Ownership category
- akta.pro rank
Prior Labs industry classification
Industry- Product category
- Tabular AI / Machine Learning Platform
- NAICS
- Computer Systems Design Services (541512), Computer Systems Design and Related Services (54151)
- SIC
- Services-Computer Programming Services (7371), Services-Prepackaged Software (7372)
- akta.pro primary industry
- Foundation Model Developers (LLM/Multimodal Model Labs) (HDAAACAA)
Keywords
Where Prior Labs is headquartered
LocationHeadquarters
- HQ city
- Berlin
- HQ country
- Germany
- HQ region
- Europe
Offices4 records
Markets served
Prior Labs business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Technology or R&D, Personnel, Infrastructure, Marketing or Sales, Operations
Revenue model
- Managed API / Inference Service: TabPFN-as-a-service via REST API or Python SDK; customers access TabPFN models on Prior Labs GPUs; pay-as-you-go consumption.
- Enterprise Licensing: Enterprise license for TabPFN-3 including on-prem and VPC deployments (AWS SageMaker, Azure AI Foundry, Google Vertex AI, Databricks, Snowflake); includes exclusive Thinking mode API; multi-year contracts.
- Cloud Marketplace Listings: TabPFN-2.5 available on AWS SageMaker Marketplace and Azure AI Foundry Marketplace for one-click deployment; transaction-fee based revenue from cloud marketplaces.
- Pre-acquisition Free Tier: Prior to SAP acquisition, Services were provided free of charge per GTC; post-acquisition transitioned to enterprise and API monetization.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Pay-as-you-go | Free Playground / Trial |
| Usage-based | Pay-as-you-go | Managed API |
| Subscription | Multi-year contract | Enterprise (TabPFN-3 Plus with Thinking mode) |
Go-to-market motion1 record
Distribution channels8 records
Marketing channels12 records
Prior Labs product offering
Product offeringCore offering
Prior Labs builds and sells the TabPFN family of pre-trained tabular foundation models — transformer-based models purpose-built for prediction on structured/tabular/relational/time-series data — delivered via managed inference API (REST + Python SDK), enterprise VPC/on-prem licenses, and a Model Context Protocol (MCP) server for LLM agents. The same model performs classification, regression, time-series forecasting, anomaly detection, and data generation in a single forward pass without task-specific training, hyperparameter tuning, or feature engineering.
Product overview
Prior Labs (Prior Labs GmbH) is a Freiburg, Germany-based AI research and product company that builds the TabPFN family of tabular foundation models — pre-trained transformers purpose-built for prediction on structured (tabular/relational/time-series) data. Its portfolio is a unified model platform rather than a suite of disconnected products: the core IP is a single training paradigm (synthetic-data pre-training for tabular inference) that is then released in successive model generations — TabPFN-v2, TabPFN-2.5, and the frontier TabPFN-3 — plus a specialized time-series variant TabPFN-TS (with TabPFN-TS-3 as a TabPFN-3 checkpoint). The same model is exposed via multiple deployment surfaces: a hosted Managed Inference API (REST + Python SDK), an enterprise VPC/on-prem offering (TabPFN-3-Plus / TabPFN Enterprise) that includes the API-exclusive Thinking mode, and a Model Context Protocol (MCP) server that turns TabPFN into a callable tool for LLM-based agents. TabPFN supports classification, regression, time-series forecasting, anomaly detection, data generation, fine-tuning, interpretability, and text integration in one foundation model, and is positioned (after the May 2026 SAP acquisition agreement) as the structured-data AI layer across the SAP Business Data Cloud and Joule agent stack, while continuing to operate as an independent research unit.
Differentiator
Problem solved
Functional benefit
Brands
- TabPFN: Prior Labs' flagship tabular foundation model family for state-of-the-art predictions on structured data. Includes model variants TabPFN-3, TabPFN-v2, and TabPFN-TS.
Products and services
- TabPFN-3 Frontier tabular foundation model released in May 2026; scaled to 1M rows and 200 features, up to 20x faster than TabPFN-2.5; supports classification, regression, time-series forecasting, many-class datasets, relational (multi-table) data, and tabular-text datasets in a single forward pass; targets enterprise data-science teams needing state-of-the-art predictions on structured data without model training or tuning.
- TabPFN-2.5 Third-generation TabPFN released November 2025; supports up to 50,000 data points and 2,000 features; substantially outperforms tuned tree-based models and matches AutoGluon 1.4 (a 4-hour tuned ensemble) on industry benchmarks; available via one-click deployment on AWS SageMaker, Azure AI Foundry, Google Vertex AI, Databricks, and Snowflake; targets enterprise customers needing cloud-marketplace-deployable tabular ML.
- TabPFN-TS (Time Series) Time-series specialized variant of TabPFN; TabPFN-TS-3 ranks 2nd on the fev-bench time-series benchmark; provides zero-shot time-series forecasting for sales, inventory, patient response, energy demand, and other temporal prediction tasks; targets enterprises needing production-grade time-series forecasting without model building.
- TabPFN Managed Inference API Hosted, GPU-backed inference API for TabPFN foundation models accessible via REST or Python SDK; provides instant access to TabPFN without requiring customer GPUs; usage-based billing; available to registered business customers for analysis, processing, and evaluation of tabular business data; targets developers and data scientists who want production-grade predictions without managing infrastructure.
- TabPFN Enterprise / TabPFN-3-Plus (Thinking) Enterprise-tier offering of TabPFN that includes the API-exclusive Thinking mode (TabPFN-3-Plus) with test-time compute scaling, on-prem and VPC deployment options (AWS SageMaker, Azure AI Foundry, Google Vertex AI, Databricks, Snowflake), enterprise licensing for large customers, and SHAP-based interpretability; targeted at Fortune-class financial institutions, healthcare systems, and industrial enterprises with strict data control and regulatory requirements.
- TabPFN MCP Server (Model Context Protocol) Model Context Protocol server that gives LLMs and AI agents (Claude, ChatGPT, others) a toolkit for classification, regression, time series, and causal inference over tables and databases via tool calls; uses streamable HTTP with OAuth or Bearer authentication; targets AI-engineering teams building LLM agents that need to make predictions on tabular data.
Quantifiable outcome
- 40% improvement in anomaly detection (Hitachi Rail)
- +13 more outcomes
Companies that use Prior Labs
Customer profileNamed customers10 records
Segments5 records
Ideal customer profiles6 records
Prior Labs technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration9 records
AI capability10 records
Feature7 records
Prior Labs partnerships and signals
Strategic signalPartnerships
Eleven partnerships are on record, tiered flagship / acquirer, core and flagship.
- SAPflagship / acquirerSAP entered into definitive agreement to acquire Prior Labs to establish a globally leading frontier AI lab in Europe. SAP will invest more than €1 billion over 4 years to scale Prior Labs into a leading AI lab for structured business data. Prior Labs will continue operating as an independent legal entity, retaining brand, team, research mission, headquarters in Freiburg, open-source strategy, and customers. Integration with SAP AI Core, SAP Business Data Cloud, and Joule agentic layer. Transaction expected to close in Q2 or Q3 2026.
- TD Bank / Layer 6 AIcoreCollaboration to explore TabPFN's application to enhance financial forecasting in lending, mortgages, and operations. TD's AI research center Layer 6 testing TabPFN for portfolio growth forecasting and risk system improvements.
- Creditplus Bank AGcorePartnership to apply TabPFN as foundation model for credit decisioning in motor financing. Prior Labs provides Creditplus with distilled models that run on CPU infrastructure, comply with credit risk regulatory requirements, and enable more accurate targeted approvals across customer segments.
- Hitachi RailcoreCollaboration with Hitachi Rail to bring TabPFN to predictive maintenance workflows across Spain's railway network. Achieved 40% improvement in RMSE compared to existing baselines; enables Hitachi to detect track anomalies 40% more accurately.
- Oxford Cancer Analytics (OXcan)coreCollaboration to apply TabPFN to high-clinical-need lung diseases and complex proteomic datasets for liquid biopsy and clinical decision-making. OXcan leverages TabPFN's proteome analysis capabilities to derive actionable clinical insights.
- TaktilecorePartnership with Taktile, a leading decision intelligence platform, to deliver TabPFN's predictive power for financial risk management applications including fraud checks, credit risk anomaly detection, and AML use cases. TabPFN ranked #1 in time series settings on GIFT-Eval benchmark via this partnership.
- DatabricksflagshipTechnical integration embedding TabPFN into Databricks Lakehouse via Unity Catalog-governed Delta tables and MLflow. Customers can run predictions on Unity Catalog data with full data governance, lineage tracking, batch and real-time inference capabilities.
- AWS (Amazon SageMaker)flagshipTabPFN-2.5 listed on AWS SageMaker Marketplace for one-click deployment into customer AWS infrastructure; supports private cloud deployment with data never leaving customer infrastructure.
- Microsoft Azure (AI Foundry)flagshipTabPFN deployed on Azure AI Foundry Model Catalog for one-click deployment on Microsoft Fabric data within Azure VPC; full data governance and security compliance.
- Google Cloud (Vertex AI)flagshipTabPFN-2.5 listed on Google Vertex AI Model Garden for one-click deployment into GCP infrastructure.
- SnowflakeflagshipEmbedded TabPFN within Snowflake Data Lake; one-click deployment for in-place predictions without data movement.
Scale indicators14 records
Recent moves7 records
Expansion highlights7 records
Prior Labs competitors and assessment
Company assessmentDirect peers
- AutoGluon (Amazon): Amazon-backed open-source AutoML framework that automates tabular ML with ensembling of gradient boosting, neural networks, and stacking. Direct competitor that TabPFN benchmarks against and is explicitly positioning to displace, including beating AutoGluon 1.4 in TabPFN-2.5 benchmarks.
- H2O.ai: Enterprise AutoML platform offering tabular prediction, feature engineering, and model interpretability for finance, healthcare, and insurance. Comparable target customers and use cases (fraud, credit risk, churn) overlap heavily with Prior Labs' enterprise segments.
- DataRobot: Enterprise AI platform with strong tabular AutoML capabilities targeted at banks, healthcare systems, and manufacturers. Same buyer personas and structured-data use cases as Prior Labs, though DataRobot is a much broader AI/ML suite rather than a single foundation model.
- RapidMiner: Established data science platform offering AutoML, predictive analytics, and ML ops for structured data with strong presence in manufacturing, finance, and healthcare. Targets the same enterprise buyer personas and tabular prediction use cases as Prior Labs.
Broad incumbents
- Google Vertex AI: Google's enterprise ML platform includes Tabular Workflows for AutoML on structured data and hosts TabPFN-2.5 in its Model Garden. Represents both a distribution channel for Prior Labs and a potential internal competitor with broader ML portfolio offerings.
- Databricks: Data lakehouse platform that offers AutoML, MLflow, and feature store capabilities for tabular data workflows. Currently an integration partner hosting embedded TabPFN, but also a potential competitor with its own AutoML and Foundation Model APIs.
- Microsoft Azure Machine Learning: Microsoft's enterprise ML platform with AutoML capabilities for tabular data, hosted on Azure AI Foundry where TabPFN-3 is also deployed. Overlaps in finance and healthcare verticals and represents both a channel and competitive threat.
- SAP Business AI / Joule: SAP's enterprise AI stack (Business Data Cloud, Joule, AI Core) is the acquirer of Prior Labs and represents the future distribution and integration context for TabPFN. Strategic parent and largest near-term customer ecosystem post-acquisition.
Emerging players
- Snowflake Cortex AI: Snowflake's in-data-cloud AI/ML service offers forecasting and ML functions on structured data inside the Data Lake. Currently hosts embedded TabPFN, but its growing native ML/AI capabilities make it both a partner and emerging competitor.
- PyCaret: Open-source low-code ML library in Python covering classification, regression, and time series on tabular data. Represents the free, easy-to-adopt alternative that competes with TabPFN for the entry-level data science audience.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat6 records
Key risks7 records
Key highlights7 records
Customer concentration
Prior Labs social profiles
Digital presencePrior Labs compliance and trust
Trust signalCompliance1 record
Prior Labs financial estimates
Financial estimateRevenue estimate
Valuation estimate
Prior Labs leadership team
Management profileNumber of profiles
Profiles5 records
Prior Labs funding detail
Funding detailFunding overview
Funding rounds1 record
Investors5 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Prior Labs 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 Prior Labs
What does Prior Labs do?
Prior Labs builds and sells the TabPFN family of pre-trained tabular foundation models — transformer-based models purpose-built for prediction on structured/tabular/relational/time-series data — delivered via managed inference API (REST + Python SDK), enterprise VPC/on-prem licenses, and a Model Context Protocol (MCP) server for LLM agents. The same model performs classification, regression, time-series forecasting, anomaly detection, and data generation in a single forward pass without task-specific training, hyperparameter tuning, or feature engineering.
Is Prior Labs a public or private company?
Prior Labs is a private company. It is classified as venture growth investor backed and is currently acquired.
When was Prior Labs founded?
Prior Labs was founded in 2024. It employs 11 to 50 people.
Where is Prior Labs based?
Prior Labs is headquartered in Berlin, Germany, in the Europe region.
How does Prior Labs make money?
Four revenue lines are on record. Managed API / Inference Service is the primary driver. The others are enterprise Licensing, cloud Marketplace Listings and pre-acquisition Free Tier.
Who are Prior Labs's main competitors?
Direct peers on record are AutoGluon (Amazon), H2O.ai, DataRobot and RapidMiner. Broad incumbents are Google Vertex AI, Databricks, Microsoft Azure Machine Learning and SAP Business AI / Joule. Emerging players are Snowflake Cortex AI and PyCaret.
Does Prior Labs have an API?
Yes. Prior Labs offers a public Managed API for TabPFN foundation models, accessible via REST API or Python SDK library, hosted on Prior Labs GPUs. It includes an API-exclusive Thinking mode (TabPFN-3-Plus). The API uses streamable HTTP with OAuth or Bearer authentication for the MCP variant and is compatible with all major LLM providers or agent builders. The API is available to registered business customers for analysis, processing and evaluation of tabular business data; personal data is not permitted as input per the Acceptable Use Policy. Developer documentation is at docs.priorlabs.ai/api-reference/getting-started.
What industry is Prior Labs in?
Prior Labs's product category is Tabular AI / Machine Learning Platform. Its primary akta.pro industry code is HDAAACAA, Foundation Model Developers (LLM/Multimodal Model Labs). Its NAICS code is 541512 and its SIC code is 7371.