Hopsworks
Hopsworks is a Stockholm-based AI infrastructure company that provides an AI Lakehouse platform with a proprietary sub-millisecond Feature Store (RonDB) for production ML systems. It serves enterprise customers in financial services, retail, government, and defense via a freemium SaaS plus enterprise deployment model.
- Company typePrivate
- Founded2018
- HeadquartersStockholm, Sweden
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What Hopsworks does
Hopsworks AB is a Stockholm-based (Hopsworks AB, headquartered at Åsögatan 119, 116 24 Stockholm) private AI infrastructure company that builds an AI Lakehouse platform centered on a Python-centric Feature Store for production machine learning systems. The platform combines an offline columnar feature store supporting open table formats (Delta, Iceberg, Hudi), an online feature store powered by proprietary RonDB technology delivering sub-millisecond retrieval (claimed 10x lower latency than SageMaker and Vertex), an integrated OpenSearch-based vector database for similarity search, and MLOps capabilities including model registry, KServe-based model serving, and FTI (feature/training/inference) pipeline orchestration via Airflow. The architecture is modular, supporting Spark, Flink, Pandas, DuckDB, TensorFlow, PyTorch, and external data warehouses (Snowflake, Databricks, Redshift, BigQuery) via External Feature Groups.
The company serves enterprise customers across financial services (Swedbank, America First Credit Union), retail and e-commerce (Zalando — powering personalization across 25 countries and 50+ million customers), gaming and betting (Paddy Power / Betfair), government and defense (Saab, Arbetsförmedlingen), education and research (Karolinska Institutet, RISE), and other industries including telecom and defense. Customers include both large enterprises and mid-market users (e.g., Clicklease), with flagship workloads such as Zalando's real-time personalization during peak Cyber Week load (millions of ops/second).
The business model combines a freemium tier (1 project, no credit card) and a usage-based Pay-As-You-Go cloud tier (run.hopsworks.ai, $0.35/compute credit) for product-led growth with self-service developers, alongside an enterprise subscription tier for bespoke deployments including on-premises and air-gapped installations. GTM combines PLG (free tier → paid conversion) with enterprise field sales targeting financial services, retail, and government, particularly in the US following a $6.5M funding round in June 2023 led by Vendep Capital. The company is ISO 27001 and SOC 2 Type II certified, GDPR and HIPAA compliant, and has been recognized via a SIGMOD 2024 research publication — the first feature store to appear at that venue.
Hopsworks firmographics
Firmographics- Name
- Hopsworks
- Legal name
- Hopsworks AB
- Website
- https://hopsworks.ai
- Company type
- Private
- Founded year
- 2018
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- Hopsworks is a Stockholm-based AI infrastructure company that provides an AI Lakehouse platform with a proprietary sub-millisecond Feature Store (RonDB) for production ML systems. It serves enterprise customers in financial services, retail, government, and defense via a freemium SaaS plus enterprise deployment model.
- Ownership category
- akta.pro rank
Hopsworks industry classification
Industry- Product category
- AI/ML Platform Software
- NAICS
- Software Publishers (5132), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518210)
- SIC
- Services-Prepackaged Software (7372), Services-Computer Processing & Data Preparation (7374), Services-Computer Integrated Systems Design (7373)
- akta.pro primary industry
- Lakehouse Platforms (HDAEABAB)
- akta.pro secondary industries
- End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management) (HDAEANAA), Enterprise AI Data & Knowledge Platforms (Vector Databases, Knowledge Graphs) (HDAEANAH), Query Engines & SQL Analytics Layers for Warehouses/Lakes (HDAEABAI)
Keywords
Where Hopsworks is headquartered
LocationHeadquarters
- HQ city
- Stockholm
- HQ country
- Sweden
- HQ region
- Europe
Offices1 record
Markets served
Hopsworks business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Infrastructure, Marketing or Sales, Operations
Revenue model
- Pay-As-You-Go Cloud Platform: Consumption-based model where customers pay only for compute credits ($0.35/credit), online storage ($0.50/GB/month), and offline storage ($0.03/GB/month). Dynamic compute with auto-scaling workers. Monthly billing via Stripe.
- Enterprise Licensing: Enterprise tier providing bespoke deployment solutions tailored to customer needs. Includes support for on-premises, air-gapped, and hybrid deployments. Contact sales for pricing.
- Freemium Tier: Free tier with 1 project, full features, and capped capacity. No credit card required. Serves as acquisition channel for conversion to paid tiers.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Monthly | Free tier: 1 project with full features and capped capacity, no credit card required |
| Usage-based | Pay-as-you-go | Pay-As-You-Go: $0.35/credit, up to 5 projects, dynamic compute with auto-scaling |
| Subscription | Annual | Enterprise: Bespoke deployment solutions tailored to customer needs, contact sales |
Go-to-market motion3 records
Distribution channels4 records
Marketing channels6 records
Hopsworks product offering
Product offeringCore offering
Hopsworks sells an AI Lakehouse platform that combines a Feature Store, an open table format data lakehouse, and MLOps capabilities for building and serving production machine learning systems. The platform provides sub-millisecond online feature retrieval through RonDB, integrated vector search via OpenSearch, model serving via KServe, and pipeline orchestration via Airflow, available as a SaaS (run.hopsworks.ai), on cloud marketplaces (AWS, Azure, GCP), or on-premises/air-gapped deployments.
Product overview
Hopsworks is an AI Lakehouse platform that provides a modular, integrated architecture combining Feature Store, AI Lakehouse, and MLOps capabilities. The core platform includes the Feature Store for managing ML feature data with sub-millisecond retrieval, the AI Lakehouse for unified data storage, and MLOps for model training, registry, and serving. Key modules include RonDB (high-performance online feature store), KServe (model serving), OpenSearch vector database (similarity search), Model Registry (versioned model storage), Airflow (pipeline orchestration), Trino (query engine), and JupyterLab (development environment). The platform supports cloud deployment on AWS, Azure, and GCP, as well as on-premises and air-gapped installations. Integrations extend to external data warehouses (Snowflake, Redshift, BigQuery, Databricks), ML training tools (SageMaker, Kubeflow, Azure ML), streaming platforms (Kafka, Flink, Spark), and enterprise tools (OAuth2, LDAP, Stripe). Security certifications include ISO 27001, SOC2 Type II, GDPR, and HIPAA.
Differentiator
Problem solved
Functional benefit
Products and services
- AI Lakehouse Central data platform combining feature store, model registry, and MLOps capabilities for building production AI systems. Provides unified storage for features, training data, and models with support for batch and real-time workloads.
- Feature Store Central repository for ML feature data with sub-millisecond retrieval latency. Supports feature reuse across models, point-in-time correct training data creation, feature monitoring for drift detection, and integration with external data warehouses (Snowflake, Redshift, BigQuery, Databricks).
- MLOps End-to-end MLOps capabilities including model registry (TensorFlow, PyTorch, Scikit-learn, LLM), model training, model serving via KServe, and inference pipeline orchestration with Airflow.
- Python API (hsfs, hsml, hopsworks) Python SDK for programmatically managing feature stores, feature groups, feature views, training datasets, model registries, model serving, projects, jobs, Kafka topics, and secrets. Available under Apache License 2.0.
- On-Premise / Air-Gapped Deployment Deployment option supporting on-premises Kubernetes clusters, including air-gapped data centers, for organizations requiring full data sovereignty. Available on Ubuntu/Redhat-compatible systems.
- Security & Compliance Package Enterprise security and compliance framework including ISO 27001 certification, SOC2 Type II compliance, GDPR compliance, HIPAA compliance, encryption at rest and in transit, multi-layered project-based access control, and audit logging.
Quantifiable outcome
- Reduced feature latency from 100ms to <1ms (Zalando)
- +3 more outcomes
Companies that use Hopsworks
Customer profileNamed customers11 records
Segments4 records
Ideal customer profiles4 records
Hopsworks technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration22 records
AI capability8 records
Feature6 records
Hopsworks partnerships and signals
Strategic signalScale indicators4 records
Recent moves6 records
Expansion highlights6 records
Hopsworks competitors and assessment
Company assessmentDirect peers
- Tecton: Tecton is a managed feature platform for production ML, originally founded by the creators of Uber's Michelangelo feature store. It is the closest direct peer to Hopsworks, offering online/offline feature serving, time-travel correctness, and enterprise MLOps integrations.
- Iguazio: Iguazio (acquired by McKinsey QuantumBlack) provided an MLOps and feature store platform with a focus on real-time AI. It was a direct peer to Hopsworks in the unified feature store/MLOps category before its acquisition.
- Feast: Feast (LF AI & Data) is the leading open-source feature store, providing online/offline feature stores with integrations to Redis, DynamoDB, BigQuery, and Snowflake. It competes with Hopsworks on the same feature store functionality and overlaps with Hopsworks' open-source positioning.
Broad incumbents
- Snowflake: Snowflake provides the AI Data Cloud with Snowpark, Iceberg tables, and ML/AI capabilities that increasingly overlap with feature store and lakehouse use cases. It is a broad incumbent that competes indirectly with Hopsworks by extending its data warehouse into ML feature serving.
- Amazon SageMaker: Amazon SageMaker is AWS's end-to-end ML platform that includes a built-in Feature Store, training, deployment, and MLOps. It is explicitly benchmarked by Hopsworks (10x latency claim) and represents the most direct hyperscaler bundling competitor.
- Databricks: Databricks offers the Databricks Lakehouse Platform including a built-in Feature Store, Unity Catalog, MLflow, and Mosaic AI. It is a broad incumbent that overlaps with Hopsworks on feature stores, lakehouse storage, and MLOps, but as part of a much larger data/AI portfolio.
- Google Vertex AI: Vertex AI is Google Cloud's unified ML platform, including Vertex AI Feature Store, model training, and serving. Like SageMaker, it is benchmarked against Hopsworks on feature retrieval latency and competes as a bundled hyperscaler alternative.
Emerging players
- Valohai: Valohai is a Finnish-origin MLOps platform providing reproducible model training, pipeline orchestration, and model management. It is a regional/niche peer to Hopsworks' MLOps module, with similar Nordic roots and target enterprise buyer.
- Weights & Biases: Weights & Biases is an MLOps platform focused on experiment tracking, model registry, and dataset versioning. It overlaps with Hopsworks' MLOps capabilities (model registry, training pipelines) but does not offer an integrated online feature store.
- Starburst (Trino): Starburst provides a commercial Trino-based distributed SQL query engine for federated analytics across data lakes and warehouses. Hopsworks bundles Trino for its offline feature queries, making Starburst a partial peer on the query engine layer of the AI Lakehouse.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat6 records
Key risks7 records
Key highlights7 records
Customer concentration
Hopsworks social profiles
Digital presenceHopsworks compliance and trust
Trust signalCompliance4 records
Hopsworks financial estimates
Financial estimateRevenue estimate
Valuation estimate
Hopsworks leadership team
Management profileNumber of profiles
Profiles9 records
Hopsworks funding detail
Funding detailFunding overview
Funding rounds3 records
Investors5 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Hopsworks 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 Hopsworks
What does Hopsworks do?
Hopsworks sells an AI Lakehouse platform that combines a Feature Store, an open table format data lakehouse, and MLOps capabilities for building and serving production machine learning systems. The platform provides sub-millisecond online feature retrieval through RonDB, integrated vector search via OpenSearch, model serving via KServe, and pipeline orchestration via Airflow, available as a SaaS (run.hopsworks.ai), on cloud marketplaces (AWS, Azure, GCP), or on-premises/air-gapped deployments.
Is Hopsworks a public or private company?
Hopsworks is a private company. It is classified as venture growth investor backed and is currently operating.
When was Hopsworks founded?
Hopsworks was founded in 2018. It employs 11 to 50 people.
Where is Hopsworks based?
Hopsworks is headquartered in Stockholm, Sweden, in the Europe region.
How does Hopsworks make money?
Three revenue lines are on record. Pay-As-You-Go Cloud Platform is the primary driver. The others are enterprise Licensing and freemium Tier.
Who are Hopsworks's main competitors?
Direct peers on record are Tecton, Iguazio and Feast. Broad incumbents are Snowflake, Amazon SageMaker, Databricks and Google Vertex AI. Emerging players are Valohai, Weights & Biases and Starburst (Trino).
Does Hopsworks have an API?
Yes. Hopsworks provides a Python API (hsfs, hsml, hopsworks) available under Apache License 2.0. The Python API enables developers to programmatically manage feature stores, feature groups, feature views, training datasets, model registries, model serving, projects, jobs, Kafka topics, secrets, and more. Users may access their data via the Application Program Interface (API) as stated in the Terms and Conditions. Developer documentation is at docs.hopsworks.ai/latest.
What industry is Hopsworks in?
Hopsworks's product category is AI/ML Platform Software. Its primary akta.pro industry code is HDAEABAB, Lakehouse Platforms, with a secondary code of HDAEANAA, End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management). Its NAICS code is 5132 and its SIC code is 7372.