Data & Memory LayerEarly StageInfraℹHyperscience
Hyperscience offers an AI software platform that turns unstructured documents into structured, machine-readable data with high accuracy.
Startup snapshot
- Stage
- Early Stage
- Funding
- Series E · $100M (Series E) · Dec 1, 2021
- Status
- published
- Updated
- Apr 6, 2026
- Views
- Category
- Data & Memory Layer
- Website
- hyperscience.com
What they're building
Hyperscience offers an AI software platform that turns unstructured documents into structured, machine-readable data with high accuracy.
Founders
Peter Brodsky
Co-founder
Foundation model usage
Editorial estimates from product descriptions. Model providers and actual usage are not confirmed.
- Inferred architecture
- Agentic loops, Batch document processing, Code generation
- Estimated model demand
- Moderate intensity
NYC footprint
- NYC office
- 285 Fulton Street, 88th Floor (One World Trade Center), New York, NY 10007
Funding
Latest funding: Series E · $100M (Series E) · Dec 1, 2021. Total raised: $299M. Lead investor: Tiger Global Management.
AI Lab Fit
Strong fit for embeddings, retrieval, extraction, document intelligence, multimodal parsing, and persistent context layers for AI applications.
Notes
Hyperscience is tagged infra based on buyer clarity, repeat workflow activity, public activity, and fit with the AI Atlas map.
Related NYC startups
Carbon Arc operates an AI-ready data marketplace, giving companies the licensed data resources they need to train more effective AI models and make better decisions.
Data & Memory LayerInfraℹRecent Activity
Wallaroo provides an enterprise AI/ML production platform that enables clients to deploy machine learning and analytics projects into production at scale, with model observability, monitoring, and inference orchestration.
Data & Memory LayerInfraℹRecent Activity
Canoe uses AI to automate alternative-investment data collection, extraction, and reconciliation for allocators, family offices, and wealth managers.
Data & Memory LayerFeaturedRecent Activity