Boltz

What Boltz does

Boltz PBC is a Cambridge, Massachusetts-based public benefit corporation that develops open-source AI foundation models and commercial software platforms for biomolecular design. The company originates from MIT CSAIL research, releasing Boltz-1 (November 2024), Boltz-2 (June 2025), and BoltzGen (October 2025) under MIT license before formally incorporating as a PBC on January 8, 2026. Its core technology spans biomolecular foundation models covering structure prediction, binding affinity estimation, and generative design across proteins, small molecules, DNA, RNA, and ligands, with experimentally validated performance benchmarks including FEP-comparable affinity accuracy at 1000x the speed, a CASP16 affinity challenge win, and a 6/10 hit-discovery rate using 28-51 compounds per target.

Its commercial product surface is built on top of these models and includes the Boltz Lab platform (an end-to-end molecular design SaaS), the Boltz API (programmatic and agent-accessible compute at $0.025 per prediction), and proprietary pipelines BoltzMol-1 (small-molecule hit discovery) and BoltzProt-1 (de novo protein and nanobody design). The company reports its models are used by over 100,000 scientists at 1,000+ companies including every top-20 global pharma, with 5M+ cumulative downloads of open-source weights.

The business model is a hybrid of open-source community distribution and paid commercial access. Revenue is generated through usage-based API fees ($0.025-$0.40 per prediction tiered by pipeline and crop size), a freemium SaaS tier (200 free monthly predictions), and custom multi-year enterprise contracts that include single-tenant or on-premise deployment, fine-tuning on customer-proprietary data, and dedicated support. Go-to-market is executed through three concurrent motions: product-led self-serve via lab.boltz.bio, API-first distribution embedded in third-party platforms (Claude Code, Codex, Gemini, Benchling, Amazon Bio Discovery, Phylo, Rowan, and others), and direct enterprise field sales targeting top-20 pharmaceutical companies, with multi-year relationships announced with Pfizer (January 2026) and Takeda (June 2026). The company has raised $28M in seed funding from Amplify Partners, a16z, and Zetta Venture Partners.

Name

Boltz

Legal name

Boltz PBC

Website

https://boltz.bio

Company type

Private

Founded year

2026

Operating status

Operating

Headcount range

1–10 employees

Short description

Boltz PBC builds open-source and commercial AI foundation models for biomolecular design—structure prediction, binding affinity estimation, and generative molecular and protein design—serving pharmaceutical, biotechnology, academic, and consumer-products researchers via SaaS platform and API.

Ownership category

akta.pro rank

Product category

AI-driven Drug Discovery / Computational Biology Software

NAICS

Research and Development in Biotechnology (except Nanobiotechnology) (541714)

SIC

Services-Commercial Physical & Biological Research (8731)

akta.pro primary industry

AI/ML Platforms for Drug Discovery & Experiment Optimization (HLAGAJAO)

akta.pro secondary industries

AI/ML-Enabled Biologic Discovery Platforms (computational design, in silico screening) (HLAAAIAE), Target Identification & Validation Platforms (omics-driven, chemoproteomics, genetic validation) (HLAAAIAI)

Keywords

• Biomolecular AI models
• Drug discovery platform
• Protein design software
• Structure prediction models
• Computational biology

Headquarters

HQ city

Boston

HQ country

United States

HQ region

North America

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Infrastructure, Personnel, Operations, Marketing or Sales

Revenue model

1. SaaS Platform Subscriptions: Usage-based SaaS platform with Professional tier (free tier with usage-based pricing) and Enterprise tier (custom pricing with unlimited GPU capacity, dedicated support, fine-tuning, and single-tenant deployment). Customers pay per prediction or molecule designed.
2. API Access Fees: Per-prediction API access starting at $0.025 per prediction. Pay-as-you-go model with no upfront costs. Small-molecule pipeline at $0.025/molecule, protein design at $0.025-$0.40 depending on crop size.
3. Enterprise Model Customization: Fine-tuning open-source models on proprietary customer data to create exclusive models for structure prediction, affinity estimation, and biologics design. Direct collaboration with Pfizer for custom model development.
4. Hosted Compute Services: High-performance compute infrastructure offered as part of platform access. Margins earned on top of compute costs through model optimization and efficient infrastructure.

Pricing tiers

Go-to-market motion3 records

Distribution channels4 records

Marketing channels6 records

Core offering

Boltz develops and commercializes AI foundation models and a SaaS platform for biomolecular design, enabling structure prediction, binding affinity estimation, and generative design of proteins, small molecules, DNA, RNA, and ligands. The company sells access through the Boltz Lab platform and Boltz API for pharmaceutical, biotech, academic, and agricultural research customers.

Product overview

Boltz is a frontier AI research lab that has evolved from MIT CSAIL into a public benefit corporation (PBC) building generative models for biology and chemistry. The product portfolio consists of two main commercial offerings—the Boltz Platform (SaaS) and Boltz API—powered by proprietary models including BoltzMol-1 (small-molecule hit discovery) and BoltzProt-1 (protein design). Underpinning the commercial products is a family of open-source models released under MIT license: Boltz-1 (structure prediction), Boltz-2 (structure + binding affinity), and BoltzGen (protein binder generation). Boltz Lab provides the user-facing platform interface with agents for automated molecular design workflows. The company also offers fine-tuning services and enterprise deployments where Boltz scientists help partners customize models on proprietary data for specific targets.

Differentiator

Problem solved

Functional benefit

Brands

• Boltz Lab: Platform for end-to-end molecular design powered by frontier models, pipelines and compute.
• Boltz API
• BoltzMol-1
• BoltzProt-1
• Boltz-2
• Boltz-1
• BoltzGen

Products and services

• Boltz Platform A SaaS molecular design platform providing end-to-end design capabilities for organizations, powered by Boltz frontier AI models, pipelines, and high-performance compute. Supports deployment via multi-tenant cloud, single-tenant, or fully on-premise with SOC 2 Type 1 certification and IP/data ownership guarantees.
• Boltz API A REST API with Python and JavaScript SDKs providing programmatic access to run Boltz AI models (BoltzMol-1, BoltzProt-1, Boltz-2) for small-molecule design, protein design, and structure prediction at scale, with pricing from $0.025 per prediction.
• BoltzMol-1 A small-molecule hit-discovery pipeline using AI co-folding for virtual screening. Identified confirmed hits on 6 of 10 challenging therapeutic targets while testing only 28-51 compounds per target. Includes structure prediction, binding affinity estimation, and ADME property prediction.
• BoltzProt-1 A protein design pipeline with de novo nanobody design capabilities. Uses Boltz-PPI (protein-protein interaction model) for binding confidence prediction, nearly tripling hit rate over BoltzGen. Delivers designs with 58% passing stringent developability panels.
• Boltz-2 An open-source biomolecular foundation model (MIT license) for joint structure and binding affinity prediction. Achieves FEP-comparable accuracy while running 1000x faster. Extended from Boltz-1 with new affinity module, improved controllability, and GPU optimizations.
• BoltzGen An open-source all-atom generative diffusion model (MIT license) for universal protein binder design. Generates nanobodies and peptides against arbitrary biomolecular targets with design specification language for constraints. Validated on 9 novel targets with 6/9 nanomolar hits.
• Boltz-1 An open-source model (MIT license) approaching AlphaFold3-level accuracy for predicting 3D structures of biomolecular complexes including proteins, nucleic acids, ligands, and their interactions. Includes training/inference code and model weights.
• Boltz Lab An early access beta platform providing user-facing interface to Boltz's AI agents and models for small-molecule hit discovery and protein design, with free monthly credits, usage-based pricing, and collaborative interfaces for computational and non-computational scientists.

Quantifiable outcome

• 6/10 confirmed hit rate with only 28-51 compounds tested per target (vs tens of thousands for traditional HTS)
• +4 more outcomes

Named customers3 records

Segments4 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration12 records

AI capability8 records

Feature8 records

Partnerships

Five partnerships are on record, tiered core.

• TakedacoreStrategic or Co-development Partner · 18 June 2026Collaboration to deploy Boltz's frontier AI models across Takeda's Research organization. Takeda scientists access BoltzMol-1 and BoltzProt-1 through Boltz Lab interface and API. Boltz scientists work on selected programs to fine-tune models for specific targets.
• BenchlingcoreTechnology or Integration · 13 May 2026Boltz models integrated into Benchling's Model Hub, enabling scientists to run Boltz models alongside experimental data without separate infrastructure. Part of Benchling's strategy to offer scientific AI models within R&D workflows.
• NVIDIAcoreTechnology or Integration · 12 January 2026NVIDIA BioNeMo platform ecosystem partner. Boltz using NVIDIA's highly optimized cuEquivariance GPU kernels for efficient molecular model inference. Part of NVIDIA's ecosystem of biotech companies developing on BioNeMo.
• PfizercoreStrategic or Co-development Partner · 8 January 2026Multi-year strategic collaboration to develop and deploy biomolecular AI foundation models. Boltz refining models on Pfizer's historical data for structure prediction, affinity estimation, and biologics design. Pfizer retains full ownership of compounds discovered. Represents flagship enterprise partnership demonstrating commercial validation.
• Amazon Bio Discovery (AWS)coreTechnology or IntegrationBoltz API natively accessible within Amazon Bio Discovery platform, providing researchers access to biological foundation models alongside AWS's AI agent for model selection and candidate evaluation.

Scale indicators8 records

Recent moves7 records

Expansion highlights7 records

Emerging players

• Profluent: AI protein design company using foundation models for de novo protein engineering. Overlaps with BoltzGen/BoltzProt-1 in generative protein design but with narrower focus on protein engineering rather than full biomolecular discovery.
• Iambic Therapeutics: AI drug discovery company using physics-informed neural networks for small-molecule design and binding prediction. Overlaps with Boltz-2 on binding affinity and BoltzMol on generative chemistry, with pipeline-internal approach.

Direct peers

• Insilico Medicine: AI-driven drug discovery company operating Chemistry42 and Biology42 platforms for generative chemistry and target discovery. Competes with Boltz across both small-molecule generation and target identification, with established pharma partnerships.
• Cradle Bio: AI protein design company offering generative models for engineering novel proteins including enzymes, antibodies, and peptides. Directly comparable to BoltzProt-1 and BoltzGen, serving biotech customers building engineered biologics.
• Isomorphic Labs: Alphabet/DeepMind spinout building frontier AI models for drug discovery including AlphaFold3. Most direct head-to-head competitor in structure prediction and biomolecular modeling; differs only in being closed-source and backed by a hyperscaler.
• XtalPi: AI-driven drug discovery platform combining quantum physics, AI, and robotics for small-molecule and crystal-form design. Competitor in AI-enabled molecular design with extensive pharma partnerships and CRO-style delivery.
• EvolutionaryScale: Frontier AI research lab building large-scale foundation models for biology, including ESM3 for protein generation. Directly comparable as a frontier-model-for-biology company with open-source releases and enterprise ambitions.
• Atomwise: AI small-molecule drug discovery company using structure-based virtual screening and proprietary AtomNet platform. Direct competitor in small-molecule hit discovery space alongside BoltzMol-1, with established pharma collaborations.
• Schrödinger: Public computational chemistry and drug discovery platform offering physics-based and AI-driven molecular simulation, FEP+ for binding affinity, and LiveDesign for collaborative discovery. Directly competes with Boltz-2 on affinity prediction and with BoltzMol on small-molecule design.
• Recursion Pharmaceuticals: AI-driven biotech combining high-throughput experimental automation with foundation models for drug discovery. Operates across small molecule and target ID, comparable to Boltz in end-to-end AI drug discovery ambition, though more integrated as a pipeline operator.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat6 records

Key risks6 records

Key highlights7 records

Customer concentration

• GitHubgithub.com/jwohlwend/boltz
• LinkedInlinkedin.com/company/boltz-bio

Compliance1 record

Revenue estimate

Valuation estimate

Number of profiles

Profiles3 records

Funding overview

Funding rounds1 record

Investors5 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Boltz