Latent Labs
Latent Labs is a London- and San Francisco-based frontier AI lab that builds proprietary all-atom generative models and an autonomous agent for de novo antibody, macrocyclic peptide, and mini-binder design, serving pharmaceutical and biotech partners through bespoke enterprise agreements and a self-serve no-code platform.
- Company typePrivate
- Founded2023
- HeadquartersLondon, United Kingdom
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What Latent Labs does
Latent Labs is a London- and San Francisco-based frontier AI lab founded in 2023 by Simon Kohl, co-developer of DeepMind's AlphaFold 2. The company builds proprietary all-atom generative AI models — Latent-X (July 2025) for macrocycles and mini-binders, Latent-X2 (December 2025) for drug-like antibodies (VHH, scFv) and macrocyclic peptides — that jointly sample protein sequence and 3D structure while modeling non-covalent interactions. Layered on top, Latent-Y (March 2026) is an autonomous AI agent that converts natural language prompts into end-to-end antibody design campaigns. These models are accessible to scientists through the Latent Labs Platform (platform.latentlabs.com), a no-code web interface with a free tier and API, and are distributed to enterprise partners through bespoke commercial agreements, private deployments, and Forward Deployed AI Engineers who embed onsite at pharma and biotech customers; a separate Research Tier for Latent-Y is scheduled to open on July 15, 2026.
The company's technology is validated in wet-lab settings rather than purely on benchmarks: Latent-X reported 91-100% hit rates for macrocycles and 10-64% for mini-binders with picomolar affinities; Latent-X2 produced hits against half of 18 tested targets using only 4-24 designs per target and is the first AI-generated antibody to clear ex vivo T-cell activation and cytokine release assays across a ten-donor human panel; Latent-Y reached a 67% target-level success rate with single-digit nanomolar affinities across nine targets and three campaign types. The stack is optimized for NVIDIA Blackwell B300 GPUs (over 4x training and 60%+ inference speedups versus prior-generation chips) and integrated with NVIDIA's BioNeMo Agent Toolkit, with additional compute relationships through AWS and Nebius and a wet-lab distribution partnership with GenScript.
Latent Labs is privately held and venture-backed. It raised a $50M seed round in February 2025 co-led by Radical Ventures and Sofinnova Partners, with participation from 8VC, Pillar VC, Kindred Capital, Isomer, Flying Fish and angels Jeff Dean (Google), Dario Amodei (Anthropic), and Mati Staniszewski (ElevenLabs). The company employs 11-50 staff across ML, applied AI, forward-deployed engineering, computational protein design, biology, and lab platform functions, and is advised by former Merck KGaA CEO Stefan Oschmann. Revenue has not been disclosed; commercial Latent-X2 and Latent-Y access is available only to 'selected partners' under bespoke agreements, while the self-serve platform operates a freemium model.
Latent Labs firmographics
Firmographics- Name
- Latent Labs
- Legal name
- Latent Labs Ltd.
- Website
- https://latentlabs.com
- Company type
- Private
- Founded year
- 2023
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- Latent Labs is a London- and San Francisco-based frontier AI lab that builds proprietary all-atom generative models and an autonomous agent for de novo antibody, macrocyclic peptide, and mini-binder design, serving pharmaceutical and biotech partners through bespoke enterprise agreements and a self-serve no-code platform.
- Ownership category
- akta.pro rank
Latent Labs industry classification
Industry- Product category
- AI-driven Computational Drug Discovery
- NAICS
- Custom Computer Programming Services (541511), Software Publishers (5132)
- SIC
- Services-Prepackaged Software (7372)
- akta.pro primary industry
- AI/ML-Enabled Biologic Discovery Platforms (computational design, in silico screening) (HLAAAIAE)
- akta.pro secondary industries
- Antibody Discovery & Engineering Platforms (display tech, bispecifics, Fc engineering) (HLAAAIAB), AI/ML Platforms for Drug Discovery & Experiment Optimization (HLAGAJAO)
Keywords
Where Latent Labs is headquartered
LocationHeadquarters
- HQ city
- London
- HQ country
- United Kingdom
- HQ region
- Europe
Offices3 records
Markets served
Latent Labs business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Infrastructure, Marketing or Sales, Operations
Revenue model
- Freemium self-serve platform: Latent Labs Platform (platform.latentlabs.com) offers a free tier with daily renewing user credits for non-coding users and academic users. Users retain a non-exclusive license to keep and use generated protein sequences; commercial use beyond the beta is permitted within the license agreement.
- Enterprise partnerships and licensing: Commercial usage of Latent-X2 and the Latent Labs Platform is available through bespoke partnership agreements with selected partners (contact: [email protected]). Includes cloud-based access and private deployments for enterprise pharma/biotech partners.
- API access: Latent-X2 and the Latent Labs Platform are accessible via API integration, enabling partners to plug model inference and binder/peptide generation directly into their own pipelines.
- Custom model deployment and forward-deployed services: Forward Deployed AI Engineers deliver end-to-end technical deployment, custom fine-tuning and integration of Latent Labs models into customer environments, including on-site work at international partner locations.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Pay-as-you-go | Free tier on Latent Labs Platform (daily renewing credits, no-code) |
| Other | Multi-year contract | Selected-partner / enterprise commercial partnership |
| Usage-based | Pay-as-you-go | API access (future/available) |
Go-to-market motion1 record
Distribution channels5 records
Marketing channels7 records
Latent Labs product offering
Product offeringCore offering
Latent Labs builds frontier all-atom generative AI models (Latent-X, Latent-X2) that jointly generate protein sequence and 3D structure for de novo design of macrocycles, mini-binders, antibodies (VHH, scFv), and macrocyclic peptides. Layered on top is Latent-Y, an autonomous AI agent that converts natural language prompts into end-to-end antibody design campaigns. Customers access these capabilities through the Latent Labs Platform (web and API) or via bespoke enterprise partnerships with forward-deployed engineers.
Product overview
Latent Labs offers a platform-plus-modules architecture for programmable biology. The unifying Latent Labs Platform (accessible via web browser and API at platform.latentlabs.com) hosts the company's frontier generative AI models and autonomous agents, providing no-code workflows for target upload, binder design, and computational ranking. The Latent-X frontier model (first generation) handles macrocycle and mini-binder design with all-atom precision, while Latent-X2 (latest generation) extends coverage to drug-like antibodies (VHH, scFv) and macrocyclic peptides with demonstrated low ex vivo immunogenicity. Latent-Y is the autonomous AI agent layered on top of Latent-X2 that orchestrates end-to-end drug design campaigns from natural language prompts. Together, these products enable partners to design de novo therapeutic molecules at scale with lab-validated results.
Differentiator
Problem solved
Functional benefit
Brands
- Latent Labs Platform: No-code web-based AI protein design platform offering access to Latent-X and Latent-X2 models via browser or API.
Products and services
- Latent Labs Platform No-code web-based AI protein design platform that hosts Latent Labs' frontier generative models (Latent-X, Latent-X2) and the Latent-Y autonomous agent. Provides scientist-friendly workflows for target upload, hotspot selection, binder design, computational ranking, and structure visualization. Accessible via web browser or API; free tier available with commercial use permitted under partnership agreements.
- Latent-X Atom-level frontier generative AI model for de novo protein binder design. Generates macrocycles (91-100% hit rates) and protein mini-binders (10-64% hit rates) with picomolar binding affinities, generating designs over 10x faster than previous methods by co-sampling sequence and structure at all-atom resolution. Available on the Latent Labs Platform with a free tier.
- Latent-X2 Frontier AI model for designing drug-like antibodies (VHH, scFv) and macrocyclic peptides with low ex vivo immunogenicity in human panels. Achieved hits on half of 18 targets with only 4 to 24 designs per target, with drug-like developability profiles emerging from the first generation without optimization, filtering, or selection. Available on the Latent Labs Platform and via API.
- Latent-Y First lab-validated autonomous AI agent for drug design. Takes natural language prompts and autonomously executes the full antibody design workflow from target analysis to lab-ready sequences, powered by Latent-X2. Demonstrated a 67% target-level success rate with single-digit nanomolar affinities across three campaign types (epitope discovery, cross-species binder design, design from publication); experts completed design campaigns 56-fold faster than working alone. Supports VHH, macrocyclic peptide, and mini-binder design.
Companies that use Latent Labs
Customer profileSegments5 records
Ideal customer profiles3 records
Latent Labs technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration1 record
AI capability9 records
Feature7 records
Latent Labs partnerships and signals
Strategic signalScale indicators12 records
Recent moves6 records
Expansion highlights5 records
Latent Labs competitors and assessment
Company assessmentDirect peers
- Isomorphic Labs: Alphabet-owned AI drug discovery company spun out of DeepMind; builds frontier computational models for drug design and has signed multi-billion-dollar pharma partnerships (e.g., Novartis, Eli Lilly). Most direct comp to Latent Labs given shared DeepMind/AlphaFold heritage and antibody/small-molecule design focus.
- Generate Biomedicines: Flagship-backed generative biology company building de novo protein therapeutics (e.g., VHHs, peptides) with a wet-lab loop; competes head-to-head with Latent Labs in AI-designed biologics and has signed large-pharma partnerships (e.g., Amgen, Pfizer).
- Absci: Public AI drug creation company combining zero-shot generative AI antibody design with high-throughput wet-lab validation. Targets the same pharma BD buyers and biologic design problem space as Latent Labs.
- Cradle: AI protein engineering platform for designing and optimizing proteins including antibodies, enzymes and peptides. Sells a SaaS-style platform to biotech and pharma users — overlapping with Latent Labs' Platform and Latent-X/X2 offerings.
- EvolutionaryScale: Frontier AI protein design lab (ESM3) founded by ex-Meta AI researchers; focuses on generative models for novel proteins and has raised large funding rounds to compete in frontier protein design alongside Latent-X/X2.
- Profluent Bio: AI protein design company building foundation models (e.g., ProteinMPNN-derived and ProGen variants) for designer proteins including antibodies and CRISPR systems; raised a $35M Series A in 2024 to scale its platform — directly comparable in stage and technology to Latent Labs.
- BigHat Biosciences: ML-guided antibody design company combining generative AI with high-throughput wet-lab characterization in a closed loop. Targets the same pharma buyers and antibody design problem space as Latent Labs' Latent-X2.
Broad incumbents
- Insilico Medicine: Public AI-driven drug discovery company with a broader pipeline across small molecules, target discovery, and biology; overlaps with Latent Labs on AI-driven therapeutic design but is significantly larger and more diversified.
- Recursion Pharmaceuticals: Public AI drug discovery platform combining high-throughput biology, imaging, and ML to discover novel therapeutics. Competes for the same large-pharma AI partnership dollars as Latent Labs but at a much broader, less antibody-focused scope.
Others
- NVIDIA BioNeMo: NVIDIA's generative AI platform and toolkit for biomolecular model training, fine-tuning, and inference — currently a partner of Latent Labs but also a potential upstream competitor offering foundation models and agents to the same pharma user base.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat5 records
Key risks6 records
Key highlights7 records
Customer concentration
Latent Labs social profiles
Digital presenceLatent Labs compliance and trust
Trust signalCompliance1 record
Latent Labs financial estimates
Financial estimateRevenue estimate
Valuation estimate
Latent Labs leadership team
Management profileNumber of profiles
Profiles15 records
Latent Labs funding detail
Funding detailFunding overview
Funding rounds3 records
Investors9 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Latent 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 Latent Labs
What does Latent Labs do?
Latent Labs builds frontier all-atom generative AI models (Latent-X, Latent-X2) that jointly generate protein sequence and 3D structure for de novo design of macrocycles, mini-binders, antibodies (VHH, scFv), and macrocyclic peptides. Layered on top is Latent-Y, an autonomous AI agent that converts natural language prompts into end-to-end antibody design campaigns. Customers access these capabilities through the Latent Labs Platform (web and API) or via bespoke enterprise partnerships with forward-deployed engineers.
Is Latent Labs a public or private company?
Latent Labs is a private company. It is classified as venture growth investor backed and is currently operating.
When was Latent Labs founded?
Latent Labs was founded in 2023. It employs 11 to 50 people.
Where is Latent Labs based?
Latent Labs is headquartered in London, United Kingdom, in the Europe region.
How does Latent Labs make money?
Four revenue lines are on record. Freemium self-serve platform is the primary driver. The others are enterprise partnerships and licensing, API access and custom model deployment and forward-deployed services.
Who are Latent Labs's main competitors?
Direct peers on record are Isomorphic Labs, Generate Biomedicines, Absci, Cradle, EvolutionaryScale, Profluent Bio and BigHat Biosciences. Broad incumbents are Insilico Medicine and Recursion Pharmaceuticals. NVIDIA BioNeMo is listed as an others.
Does Latent Labs have an API?
Yes. Partners can integrate their systems with the Latent Labs API to access the platform's generative biology models (Latent-X, Latent-X2, Latent-Y). Commercial usage is available through partnership agreements. Customers can alternatively access the platform in a web browser. Documentation and practical tips are available at docs.latentlabs.com. Developer documentation is at docs.latentlabs.com.
What industry is Latent Labs in?
Latent Labs's product category is AI-driven Computational Drug Discovery. Its primary akta.pro industry code is HLAAAIAE, AI/ML-Enabled Biologic Discovery Platforms (computational design, in silico screening), with a secondary code of HLAAAIAB, Antibody Discovery & Engineering Platforms (display tech, bispecifics, Fc engineering). Its NAICS code is 541511 and its SIC code is 7372.