Noetik
Noetik is an AI-native biotechnology company building multimodal transformer foundation models (OCTO, OCTO-vc, TARIO-2) trained on proprietary spatial tumor data to predict cancer patient treatment response, licensing the models and biomarkers to pharma (GSK, Agenus) while developing an internal therapeutics pipeline.
- Company typePrivate
- Founded2023
- HeadquartersSouth San Francisco, United States
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What Noetik does
Noetik is an AI-native biotechnology company founded in 2023 by Ron Alfa (former SVP, Head of Research at Recursion) and Jacob Rinaldi, headquartered in South San Francisco. The company's core platform integrates three layers: (1) an in-house Human Data factory that generates paired multimodal spatial omics measurements (H&E, multiplex protein panels, whole spatial transcriptomes, and whole exome sequencing) from tumor resections using patent-pending tissue microarray (TMA) processes, currently at 2.9PB of data, 450M cells, 196,000 cores, and 4,500+ patients; (2) a family of foundation models trained on that data, including the OCTO "world model," OCTO-VirtualCell (OCTO-vc) virtual cell simulator, and TARIO-2 response predictor; and (3) downstream applications such as the Perturb-map in vivo functional genomics platform and the Celleporter interactive visualization tool.
The technology uses a custom structured multimodal masking objective on spatially aligned tokens across modalities, with the OCTO base model pre-trained on 20B tokens across 128 GPUs and the OCTO-vc model reaching 1.5B parameters on approximately 200M cells. TARIO-2 uniquely infers spatial transcriptomic-grade response prediction from routine pretreatment H&E slides alone, which is the primary commercial artifact.
Noetik monetizes the platform through multi-year, non-exclusive enterprise licensing of foundation models to pharma (anchor reference: a five-year, $50M+ upfront deal with GSK for OCTO-VC in NSCLC and colorectal cancer) and strategic co-development collaborations with biotech (e.g., Agenus for predictive biomarkers on BOT/BAL immuno-oncology). Pricing is bespoke and quote-based; no self-serve or marketplace exists. A pre-revenue internal cancer therapeutics pipeline is also being built out under an EVP of Therapeutics. The company has approximately 30 employees as of mid-2025 and is backed by DCVC, Polaris Partners, Khosla Ventures, and a syndicate of life-sciences-focused investors following a $14M seed (September 2023) and $40M Series A (August 2024).
Noetik firmographics
Firmographics- Name
- Noetik
- Legal name
- Noetik Inc.
- Website
- https://noetik.ai
- Company type
- Private
- Founded year
- 2023
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- Noetik is an AI-native biotechnology company building multimodal transformer foundation models (OCTO, OCTO-vc, TARIO-2) trained on proprietary spatial tumor data to predict cancer patient treatment response, licensing the models and biomarkers to pharma (GSK, Agenus) while developing an internal therapeutics pipeline.
- Ownership category
- akta.pro rank
Where Noetik is headquartered
LocationHeadquarters
- HQ city
- South San Francisco
- HQ country
- United States
- HQ region
- North America
Offices1 record
Markets served
Noetik business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Operations, Supply Chain, Infrastructure, Marketing or Sales
Revenue model
- Foundation model licensing to pharma: Multi-year non-exclusive licensing of AI foundation models (e.g., five-year, $50M upfront plus near-term milestones deal with GSK for OCTO-VC models in NSCLC and colorectal cancer). Revenue recognized upfront and through milestone payments tied to research and commercial progress.
- Research collaboration and biomarker development: Strategic collaborations with biotech/pharma (e.g., Agenus) to develop AI-enabled predictive biomarkers for specific therapeutic programs, with exclusive rights for the partner to apply the resulting biomarker insights in drug development and commercialization.
- Internal therapeutics pipeline (pre-revenue): Internal cancer therapeutics pipeline, led by EVP of Therapeutics, advancing programs from AI-driven discovery toward early clinical development. Not yet generating product revenue.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Other | Multi-year contract | Enterprise licensing of foundation AI models (e.g., OCTO-VC virtual cell models for NSCLC and colorectal cancer) to pharma partners via multi-year, non-exclusive licenses with upfront payments and milestone-based economics. |
Go-to-market motion3 records
Distribution channels4 records
Marketing channels7 records
Noetik product offering
Product offeringCore offering
Noetik develops and licenses AI foundation models for oncology drug discovery, trained on proprietary in-house generated multimodal spatial omics datasets from thousands of patient tumors. Its core products (OCTO, OCTO-vc, TARIO-2) enable prediction of cancer patient treatment response, identification of novel therapeutic targets through counterfactual virtual cell simulations, and image-based patient stratification from routine H&E pathology slides. The company monetizes these models through multi-year enterprise licensing agreements with pharmaceutical and biotechnology companies, supported by an integrated wet-lab data factory and in vivo Perturb-map validation platform.
Product overview
Noetik operates an integrated AI-native drug discovery platform consisting of three interconnected pillars: (1) a proprietary Human Data layer that generates large-scale, spatially aligned multimodal tumor datasets (H&E, multiplex protein, spatial transcriptomics, WES) from thousands of patients in-house; (2) a suite of Foundation Models trained on that data — the OCTO world model, the OCTO-vc (virtual cell) model, the TARIO-2 therapeutic response predictor, and the Celleporter visualization interface; and (3) Applications that translate model outputs into concrete drug development outcomes, including the Perturb-map in vivo functional genomics platform for target validation and biomarker discovery. The platform is monetized both by out-licensing foundation models to pharma (e.g., the GSK OCTO-VC deal) and through research collaborations (e.g., Agenus BOT/BAL biomarker work).
Differentiator
Problem solved
Functional benefit
Brands
- OCTO: Oncology Counterfactual Therapeutics Oracle — Noetik's first foundation model (world model) for cancer biology, a multimodal transformer-based model trained on spatially aligned multiplex protein, gene expression, DNA sequencing, and H&E data.
- OCTO-VirtualCell (OCTO-VC)
- Celleporter
- TARIO-2
- Perturb-map
Products and services
- OCTO (Oncology Counterfactual Therapeutics Oracle) Noetik's first-in-class world model for cancer biology: a massively multimodal transformer pretrained on spatially aligned H&E, multiplex protein, spatial transcriptomics, and WES data via structured multimodal masking. Capable of simulating 16-channel protein images from heavily masked prompts and running counterfactual perturbation simulations to identify drug targets and predict therapeutic response. Designed to be licensed to pharmaceutical R&D organizations for indication-specific use.
- OCTO-VirtualCell (OCTO-vc) Next-generation virtual cell foundation model trained on ~40M cells (later scaled to ~200M) across many cancer indications. Simulates spatially resolved single-cell gene expression at arbitrary positions in real patient tissue, enabling high-fidelity prediction of thousands of genes in response to changes in biological context. Used to stratify patient populations and identify therapy resistance mechanisms (e.g., KRAS/STK11). Licensed to GSK under a five-year, $50M+ upfront non-exclusive license for NSCLC and CRC research.
- TARIO-2 Multimodal AI foundation model that predicts patient response and survival outcomes to next-generation immune checkpoint inhibitor combinations (e.g., Agenus's BOT+BAL) using only routine pretreatment H&E pathology images. Trained on thousands of samples across 19 cancer types and paired with high-resolution spatial transcriptomics during training. Outperformed leading pathology AI tools in retrospective MSS mCRC validation (64% vs. 9% response rate; HR 0.18 for OS).
- Perturb-map High-throughput in vivo perturbation platform that introduces hundreds of tumor subpopulations with unique mutations (suggested by Noetik's models) into a single mouse, then spatially resolves their impact on the tumor microenvironment. Used to validate AI-generated therapeutic hypotheses, study tumor-immune interactions, and map how mutations affect response to immune checkpoint blockade. Initial dataset maps >600 perturbations on tumor immunotypes.
Quantifiable outcome
- 64% response rate vs 9% in remaining cohort for BOT+BAL in MSS mCRC (no active liver metastases), with overall survival hazard ratio of 0.18, in a retrospective study of 113 patients using TARIO-2 on routine pretreatment H&E images.
- +4 more outcomes
Companies that use Noetik
Customer profileNamed customers2 records
Segments3 records
Ideal customer profiles3 records
Noetik technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- No
- API docs
- API detail
Core technology
AI maturity
App detail
Integration1 record
AI capability8 records
Feature8 records
Noetik partnerships and signals
Strategic signalPartnerships
Four partnerships are on record, tiered core, flagship / anchor partnership and core / flagship.
- Bruker Spatial BiologycoreExpansion of an existing collaboration to advance AI-driven tissue foundational models for translational and therapeutic applications in oncology. Builds on a prior study of more than 3,500 patient samples using Bruker's CosMx Spatial Molecular Imager, with Noetik now imaging thousands of additional patients to train its models toward one billion spatially resolved human cells. Supports Noetik's OCTO series and Oncology Counterfactual Therapeutics Oracle virtual cell models.
- GSKflagship / anchor partnershipFive-year strategic collaboration and non-exclusive licensing agreement granting GSK access to Noetik's OCTO-VC virtual cell foundation models for non-small cell lung cancer (NSCLC) and colorectal cancer research. Deal valued at $50M upfront plus near-term milestones. Combines GSK's tumor immunology expertise with Noetik's spatial biology platform. Described as the first substantial pure-play AI licensing deal in the TechBio sector and a potential template for AI infrastructure licensing in biopharma. Noetik retains the right to license OCTO-VC to other drug developers.
- Agenuscore / flagshipResearch collaboration to develop AI-enabled predictive biomarkers for Agenus's botensilimab (BOT) and balstilimab (BAL) immuno-oncology treatments using Noetik's OCTO virtual cell foundation model (1.5B parameters, trained on multimodal spatial data from nearly 200M tumor and immune cells). Spans colorectal, lung, ovarian, and sarcoma cancers. Agenus receives exclusive rights to apply resulting biomarker insights in drug development and commercialization. Resulted in ASCO 2026 retrospective data showing 64% response rate vs 9% in TARIO-2-identified BOT+BAL subgroup in MSS mCRC without active liver metastases.
- Khosla Ventures
Scale indicators11 records
Recent moves6 records
Expansion highlights6 records
Noetik competitors and assessment
Company assessmentDirect peers
- Recursion Pharmaceuticals: AI-native biotechnology company combining high-throughput biology with foundation-model-style ML to accelerate drug discovery. Noetik's two co-founders (Ron Alfa and Jacob Rinaldi) are both former Recursion leaders, making it the closest conceptual and operational analogue in the public TechBio space.
- Insitro: ML-driven drug discovery company combining induced pluripotent stem cell biology with foundation models and large-scale data generation. Overlaps with Noetik in using proprietary wet-lab data to train ML models for oncology and other indications, and competes for similar pharma partnerships and talent.
- Generate Biomedicines: Pioneer of generative biology foundation models that learn the language of proteins and apply it to therapeutic design. Directly comparable to Noetik's OCTO family of multimodal foundation models, and similarly targets pharma partnerships and internal therapeutics programs.
- Owkin: AI-for-oncology and biopharma company that builds foundation models on multimodal patient data (pathology, omics, clinical) and partners with pharma on biomarker, stratification, and trial-enrichment use cases. Closely parallels Noetik's TARIO-2 thesis and pharma-licensing GTM.
- Exscientia: AI-driven precision medicine company that combines generative chemistry with patient-derived data for drug design and selection. Comparable to Noetik in selling AI capabilities to large pharma partners and using data-driven approaches to oncology drug development.
- PathAI: AI-pathology company that builds foundation models on histopathology images for biomarker discovery, patient stratification, and pharma partnerships. Closely mirrors Noetik's TARIO-2 image-based response prediction approach and pathology-AI focus.
- Absci: Generative AI drug creation company combining foundation models for proteins with a high-throughput wet lab. Overlaps with Noetik in building proprietary data + foundation models to enable new therapeutics, and in pursuing both internal pipeline and pharma collaborations.
- BenevolentAI: AI-drug discovery company that builds knowledge graphs and ML models to identify novel targets and indications, including in oncology. Comparable in combining AI/ML with pharma partnerships and pursuing precision-medicine applications in cancer.
Broad incumbents
- Tempus Labs: Large, established AI-powered precision medicine company with massive multimodal oncology data (DNA, RNA, imaging, clinical) and extensive pharma services. A broad incumbent in the same oncology AI/data space Noetik targets, with a much larger scale and broader product portfolio.
Emerging players
- Lantern Pharma: AI-driven oncology company using machine learning on genomic and clinical data to develop and reposition cancer therapies, with an emphasis on response prediction. Less capital-intensive than Noetik but pursues a similar AI-for-oncology-therapeutics thesis.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat6 records
Key risks7 records
Key highlights7 records
Customer concentration
Noetik social profiles
Digital presenceNoetik financial estimates
Financial estimateRevenue estimate
Valuation estimate
Noetik leadership team
Management profileNumber of profiles
Profiles13 records
Noetik funding detail
Funding detailFunding overview
Funding rounds3 records
Investors16 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Noetik 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 Noetik
What does Noetik do?
Noetik develops and licenses AI foundation models for oncology drug discovery, trained on proprietary in-house generated multimodal spatial omics datasets from thousands of patient tumors. Its core products (OCTO, OCTO-vc, TARIO-2) enable prediction of cancer patient treatment response, identification of novel therapeutic targets through counterfactual virtual cell simulations, and image-based patient stratification from routine H&E pathology slides. The company monetizes these models through multi-year enterprise licensing agreements with pharmaceutical and biotechnology companies, supported by an integrated wet-lab data factory and in vivo Perturb-map validation platform.
Is Noetik a public or private company?
Noetik is a private company. It is classified as venture growth investor backed and is currently operating.
When was Noetik founded?
Noetik was founded in 2023. It employs 11 to 50 people.
Where is Noetik based?
Noetik is headquartered in South San Francisco, United States, in the North America region.
How does Noetik make money?
Three revenue lines are on record. Foundation model licensing to pharma is the primary driver. The others are research collaboration and biomarker development and internal therapeutics pipeline (pre-revenue).
Who are Noetik's main competitors?
Direct peers on record are Recursion Pharmaceuticals, Insitro, Generate Biomedicines, Owkin, Exscientia, PathAI, Absci and BenevolentAI. Tempus Labs is listed as a broad incumbent. Lantern Pharma is listed as an emerging player.
Does Noetik have an API?
No public API is recorded for Noetik.