Search
Contact salesLog in to Token FactoryLog in to AI Cloud

AI Discovery Awards

AI Discovery by Nebius is an annual awards for startups that are leveraging AI to revolutionize biopharma, genomics, digital health, medical devices, and medical imaging.

Our mission is to spotlight new technologies, foster collaboration between startups and potential investors, and propel the next generation of breakthroughs in healthcare.

See the list of semifinalists

Awards categories

BioPharma

Accelerating drug discovery with computational biology, molecular modeling, and AI-driven compound design.

Genomics

Advancing sequencing analysis with scalable workflows for variant detection and genomic insights.

Digital Health

Transforming clinical workflows with AI-powered patient interaction and real-world data intelligence.

Medical Devices

Enabling real-time AI for connected medical equipment and intelligent bedside systems.

Medical Imaging

Enhancing radiology and pathology with AI-driven image analysis and multimodal diagnostics.

Semifinalists

Out of 646 submissions, 198 have advanced to the semifinal following the first round of evaluations. Finalists will be announced in June!

Aikya.bio

Aikya is an always-on, multi-agent evidence generation infrastructure for antimicrobial resistance drug discovery. The platform utilizes a specialized multi-agent architecture where transformer LLMs are fine-tuned per pathogen organism, e. g. K. pneumoniae and S. aureus.

Athos Omics AI

Agentic AI platform that enables life science companies/researchers to quickly discover insights from their omics data using data, AI and scientific libraries.

Atombeat, Inc

Hermite Drug Design Software is a platform with advanced analytics tools powered by pre-trained models for drug discovery workflows. It supports target validation, hit discovery, lead optimization, and ADME/T analysis across small molecules, peptides, antibody-drug conjugates, and radionuclide drug conjugates.

Atomistic Insights

DeepPath is Atomistic Insights’ physics guided AI platform for predicting protein dynamics in drug discovery. DeepPath generates realistic conformational ensembles and transition pathways, helping discovery teams identify hidden binding pockets, understand disease relevant protein motion, and prioritize better drug candidates.

BluMaiden Biosciences

BluMaiden, in collaboration with Cedars-Sinai, is pursuing a first-in-class small molecule therapeutic for intestinal fibrosis — a severe and irreversible complication of Crohn's disease for which there is currently no approved treatment. The drug candidate is human-derived and biocompatible by design, meaning it is built from molecules already present in human biology and refined by evolution for safety and specificity.

Chemetrian

Chemetrian is an AI co-pilot that compresses months of experimental work into weeks by intelligently guiding chemists through every stage of molecular discovery and process optimization. The platform integrates four AI-powered capabilities that work together to dramatically reduce experimental burden.

ChemIllusion

ChemIllusion is an AI-native chemistry workspace for creating, validating, and communicating molecular information. It combines a browser-based chemical drawing tool with AI-powered generation workflows so users can move from molecular input to chemically grounded images, figures, graphical abstracts, and slide content in one environment.

clock.health

imAgeScore is an image-based cellular aging platform that quantifies the biological age of human cells and enables high-throughput discovery of rejuvenation interventions. Built on Cell Painting and machine learning, imAgeScore analyzes high-content microscopy features from primary human dermal fibroblasts to predict phenotypic cellular age.

CuragenAI

CureRX is an AI-native drug discovery platform that integrates generative chemistry, structure-based reinforcement learning, and multi-objective ADMET optimization to design novel small molecule therapeutics. The platform uses a graph-based transformer architecture for scaffold-constrained de novo drug design, trained via three-phase transfer learning.

CytoReason

CytoReason’s Scientific Portfolio Management (SPM) platform leverages computational disease modeling to transform vast multimodal biomedical data into scalable, interpretable insights on the relationships between diseases, gene targets, drugs, and patient populations. It enables data-driven decision-making across the drug development lifecycle, supporting key applications such as target prioritization and indication expansion.

Decoy Therapeutics

D-MAVs (Designable Multi-Antivirals) are a new category of peptide antivirals engineered to target broadly conserved viral fusion mechanisms rather than specific viral proteins. D-MAVs function as "decoys" by anchoring into the host cell membrane and binding to conserved regions of the viral fusion machinery, effectively blocking infection at the earliest possible stage of cell entry.

Deep Science Ventures

Elman is verifiable AI for drug discovery, an autonomous discovery engine for the life sciences. Given a disease and a research question, it ingests public omics data and primary literature into a unified knowledge graph at arm-level resolution, generates and verifies candidate mechanistic hypotheses, consolidates the survivors into intervention points, and predicts cohort response through an integrated trial-simulation module.

deepmirror

Elio is the foundation model for molecule design, combining a constantly updated pool of curated data with customer-specific proprietary datasets to create superior AI models for optimal drug design. This is the most complete model for molecular properties and enables chemists to deploy AI directly into their research workflows, with faster iteration, more relevant predictions, and better outcomes.

Denovo Sciences

Denovo Platform is a proprietary end-to-end drug discovery system that combines reinforcement learning, generative AI, cheminformatics, and physics-based molecular simulations to design novel small-molecule drug candidates. The platform learns from simulation-generated feedback, enabling discovery for challenging, novel, or data-poor biological targets.

DIOSynVax

DIOS Vaccines designs are based upon a foundational protein language model which is then fine tuned to represent the evolution of a viral family of interest. Because the model is trained on a vast amount of viral sequences, it generates novel sequences that cover a wide range of viral families. For this reason, the optimised sequences can be broadly effective on many different strains of a particular infectious disease.

Elnora

Elnora is an AI agent that generates and optimizes preclinical lab protocols for biotech and pharma teams. She is reachable in the terminal via CLI and MCP, easily integrated with other agents and products via a comprehensive API layer, and also has an intuitive web UI.

Embryonet AI

The proprietary metric learning platform analyzes dynamic developmental processes in zebrafish embryos and human organoids, positioning drugs in multidimensional mechanism space. Rather than asking "what drug is this?" we answer "where does this drug sit in the mechanism landscape?".

Envalyx

BEAST is an enzyme design platform for protein and multienzyme system engineering. It combines protein language models with structure prediction, molecular docking, simulation, sequence design, and computational validation to rank enzyme candidates before wet-lab testing.

Explority AI

Explority AI has trained LLMs to predict the probability of success for early-stage therapies across rare diseases using published research data. Its platform and personalized reports help pharma teams identify high-potential assets for deal sourcing and R&D prioritization.

Gardn Biosciences

GARDN Bio develops a software platform for designing RNA therapeutics for the pharmaceutical industry. Its proprietary models and inference technology optimize protein expression, tissue targeting, durability, and immune response simultaneously to improve the likelihood of clinical success.

Gero

ProtoBind is a sequence-first generative AI platform that bypasses 3D-structural bottlenecks to accelerate hit discovery for “undruggable” targets, delivering hit-finding precision comparable to leading structure-based solutions.

Harell Data Corp

Harell Data is building infrastructure for scientific discovery by combining secure proprietary data sharing with the compute and orchestration layers needed to train large-scale models. Its integrated platform brings compute to the data, enabling the development of domain-specific foundation models from scientific datasets.

Hummingbird Bioscience

Hummingbird Bioscience’s AI-first workflow uses AI to identify therapeutic epitopes and design immunogens, while the Abvantage immune-engineered mouse generates high-quality antibody sequences through in vivo affinity maturation. Computational tools then optimize and narrow those sequences into a small set of drug-like candidates for testing.

Infiltrate Bio

The Infiltrate Bio Platform is a biomimetic chip and predictive AI layer that ranks therapeutic cell products before they enter a patient or an in vivo study. Each experimental run captures up to millions of single-cell biophysical measurements under physiological flow, including per-cell velocity, rolling behavior, and arrest events.

JoeLuT AI

BenchVoice is a voice-first electronic lab notebook built for researchers. Instead of stopping an experiment to type notes, researchers speak. BenchVoice records, transcribes, and automatically extracts reagents, quantities, observations, and experimental parameters into structured lab reports.

LynxKite

LynxKite develops an AI orchestration and model-training platform for preclinical drug discovery, with an initial focus on computational longevity and age-related diseases. The product combines biomedical knowledge graphs, graph neural networks, virtual screening, molecular modelling, molecular dynamics validation, and GPU-scale workflow orchestration in one integrated environment.

Molecular Glue Labs

Cherry is a foundation model for protein–protein interaction physics trained on quantitative measurements of protein interactions inside living cells rather than evolutionary sequence data. The model learns binding energetics, residue-level effects, and cooperative interaction patterns that emerge in cellular environments.

NEBULA

NEBULA has developed a physics-driven generative AI platform that combines molecular dynamics simulations with reinforcement learning. Its dataset-free approach autonomously explores biologically relevant protein conformations from first principles, enabling unbiased mapping of protein dynamics.

NenulAi

CURE.SIM is an in silico cancer simulation platform that models the tumor microenvironment using multimodal data to recreate patient-relevant tumor biology. It generates predictive models of tumor evolution, resistance, dormancy, and recurrence to help identify promising therapeutic candidates before clinical testing.

Neoncorte Bio

Neoncorte AI Sequence Optimization Platform optimizes proteins, enzymes, and aptamers using customer experimental data, including mutation–fitness tables, binding curves, activity measurements, and SELEX sequencing data. It applies protein language models, inverse folding, and custom aptamer architectures to model and optimize sequence space.

Nerai Bioscience

Nerai Bioscience is developing PAM-specific CRISPR base editors to treat rare liver metabolic diseases beyond the reach of current genome editing technologies. By correcting genetic defects directly in hepatocytes in vivo, the approach aims to restore liver function and reduce the need for lifelong care or liver transplants.

Nexus Concordat

AETHER is a transformer-based language model architecture that replaces standard softmax attention with Hodgkin-Huxley conductance gating driven by a persistent 9-dimensional neurochemical state vector (dopamine, serotonin, oxytocin, cortisol, norepinephrine, adrenaline, GABA, vasopressin, prolactin).

Om Therapeutics

OmPLex adapts protein language models and contrastive learning for highly-accurate drug-target interaction predictions. Consisting of a pre-trained protein language model, circular fingerprints for small molecule drugs, and triplet-based contrastive loss methods for drug-target-decoy updated learning.

One Zyme

One Zyme helps biopharma teams identify licensing and acquisition opportunities, monitor competitors and pipelines, and detect commercial buying triggers such as clinical trials, financings, and regulatory milestones. The platform enables faster evaluation, competitive intelligence, and more targeted business development outreach.

Optimol

Optimol Freegen is an AI-enabled drug discovery platform combines a proprietary free energy framework with machine learning to enable rapid and accurate prediction of molecular binding affinities.

Orbion

Orbion is an AI software platform for protein engineering and experimental planning, designed for biotech, pharmaceutical, and CRO teams working on difficult protein targets. The platform helps scientists make earlier and better decisions across several parts of the workflow that are still often handled through fragmented tools, manual iteration, and trial-and-error.

Owkin

K-Pro is an AI Scientist enabling smarter and faster pipeline decisions. K Pro generates hypotheses, validates them experimentally, and continuously refines them using real patient data through Owkin’s patient validation infrastructure. This creates a closed-loop learning system that translates scientific insights into real-world impact.

Peptris Technologies

PepAI is a proprietary AI algorithms minimize the number of compounds requiring wet-lab testing, accelerating and reducing the cost of drug discovery. Its foundational unsupervised models learn from large compound libraries to accurately predict compound-target activity and physicochemical properties.

Phylo

Biomni is the first general-purpose biomedical AI agent. Biomni Lab is an Integrated Biology Environment providing a unified workspace where biologists can collaborate directly with AI agents to analyze complex datasets, query literature, design molecules, generate experimental protocols, and launch wet-lab experiments.

Phytovenomics

Proteus is an autonomous biologics design platform that runs end-to-end protein design campaigns from a user-provided target. It analyzes target surfaces, generates candidates across multiple modalities, scores and filters them for binding, stability, and developability, and iteratively redesigns failed candidates.

Pierogi Therapeutics, Inc

The company is developing the first ML model to predict astatine chemistry for At-211 radiopharmaceutical development. By modeling the behavior of this rare radioisotope, the platform aims to reduce deastatination, a major failure mode in targeted alpha therapy clinical trials.

QuantaBricks LLC

ChemOrchestra is an orchestration platform for AI-driven drug discovery and molecular design. It integrates biological simulation, molecular modeling, and multi-model reasoning workflows to help researchers prioritize compounds, compare hypotheses, and accelerate experimental decision-making.

Quantum-Clarity LLC

Prometheus is an AI-enabled scientific decision platform for computational drug discovery and biopharma research. It helps scientists extract actionable insights from molecular simulations and other computation-heavy workflows by improving interpretation, prioritization, and reporting.

Rayca Precision

The Rayca Discovery Engine is an end-to-end, agentic AI + physics platform that designs, simulates, validates, and learns from the wet lab in one closed loop, generating partnership-ready therapeutic assets on repeat.

ReasonBio, Inc

ReasonBio builds reasoning foundation models for molecular design. The flagship model, Rebio Preview, is the first protein design system that performs deliberative reasoning before generation, applying the same RL-trained reasoning paradigm behind models like o1 and Claude to biology.

Swix

Swix is a software platform for AI‑driven biological simulation, built around modular, tunable building blocks that compose into multi‑scale models — from molecular forces and conformational sampling, through subcellular interactions and whole‑cell behavior, toward organism‑level dynamics. Machine learning is integrated end to end and execution is GPU‑native at every layer.

Sygnature Discovery

Sygnature Discovery proposes a novel framework that leverages the familiar Electronic Lab Notebook (ELN) paradigm. By formalizing AI workflows as ELN protocols and treating AI execution as ELN experiments, the proposed system provides a scalable, traceable, and user-oriented deployment strategy that aligns with existing laboratory practices.

Synlico Inc.

Synlico’s product is CausalCell AI, a patient-native causal AI platform that converts large-scale primary patient single-cell data into a digital causal blueprint of disease biology. The platform is designed to answer the core therapeutic question: if we intervene on a gene, pathway, or drug input in a specific patient context, what functional change occurs downstream?

Synthyra

Atlas is a high-speed AI platform that predicts protein–protein interactions directly from amino acid sequences, without requiring 3D structures. It enables pharma and biotech teams to rapidly screen for off-target effects, map host–pathogen interactions, and evaluate biologic candidates at a scale and accuracy that previous methods could not achieve.

Tahoe Therapeutics

Tahoe-100M is a giga-scale open-source single-cell perturbation atlas containing over 100 million transcriptomic profiles across thousands of drug–cell interactions. Built as foundational training data for next-generation AI “virtual cell” models, it is far larger than previously available public drug-perturbed single-cell datasets.

Technetium Therapeutics

TC-43 is an agentic AI drug design engine developed by Technetium Therapeutics on its TechnetiumX platform. Instead of using a linear screening pipeline, it uses cooperating AI agents in a closed-loop system to iteratively design, evaluate, and refine small-molecule drug candidates against therapeutic targets.

Tekkare

Tekkare’s OIP platform is a healthcare intelligence and verification system that unifies hundreds of official healthcare data sources — including clinical trials, publications, patents, regulatory approvals, OMICS, and epidemiology — into a single searchable infrastructure for pharma and research teams.

Theraimmune

Theraimmune is a multi-modal drug discovery platform specifically designed to identify and de-risk therapeutic targets for menopause-related, female-focused age-related diseases.

Vantro Health

Vantro Health is a SaaS market entry platform for biopharma, medtech, and biotech companies entering global markets. The platform serves BD, market access, health economics, and regulatory affairs teams at pharma and biotech companies — from mid-size innovators to global enterprises — who need faster, more structured market entry decisions.

Virgo

EndoDINO is an AI foundation model for endoscopic imaging trained on one of the largest endoscopy video datasets reported in the literature. It achieves state-of-the-art performance across GI endoscopy benchmarks and can predict ulcerative colitis treatment response from baseline colonoscopy videos.

VUGENE

VUGENE is an AI-powered platform that analyzes multi-omics data — including genomics, transcriptomics, proteomics, and metabolomics — within a unified framework to decode drug effects on biological systems. It reconstructs biological pathways from target engagement to downstream effects, transforming raw experimental data into decision-ready insights.

Xaira Therapeutics

X-Cell is a virtual cell AI model that predicts cellular responses to genetic perturbations across the genome without physical experiments. It uses a diffusion language model that integrates multimodal biological knowledge, including protein interaction networks, cancer dependency maps, and protein language models, into its generative process.

Xelari Inc.

Xelari Platform is a computational system for designing functional DNA/RNA binders (Xelamers, the same functional class as aptamers, also called "chemical antibodies") entirely in silico. It replaces traditional SELEX with a GPU-accelerated pipeline that moves from target protein to validated binder in up to 24 hours.

Preliminary timeline

May 8, 2026

Applications submission deadline

May 20, 2026

Announcement of the semifinalists

June, 2026

Final consideration

July 1, 2026

Official award ceremony in London

Winners receive

GPU cloud credits

1st place $100,000, 2nd place $50,000 and 3rd place $30,000.

Media exposure and promotion

Benefit from media exposure and digital promotion.

Introduction to our friendly VCs

Pitch your idea to potential investors from our trusted network.

Mentorship session

Verify ideas and get valuable insights with Nebius C-level management.

Winner trophies

Get exclusive winner’s trophy to celebrate your success.

Prizes for everyone

All participants receive a SWAG package and free inference credits for Token Factory.

Who can apply

Each participant should have a legal entity, a functioning website, product MVP and GTM strategy.

Startups developing AI products in fields such as biopharma, genomics, medical devices, medical imaging.

Bootstrapped, self-funded and funded startups at any stage from pre-seed up to and including Series D.

Qualification criteria

AI-powered

The product leverages advanced AI/ML models that require scalable GPU infrastructure for training, fine-tuning or inference, effectively utilizing high-performance computing environments.

Technological innovation

The product or service advances the industry by introducing new capabilities or significant improvements.

Functionality and advantages

The product offers superior or unique functionality, providing a competitive edge and enhancing business processes.

Performance and efficiency

The product delivers measurable results with clear metrics or case studies, improving customer performance effectively.

Global impact

The product addresses unique needs and solves specific problems, creating meaningful value on a global scale.

Market potential and business sustainability

The product demonstrates strong market demand and long-term potential, ensuring growth opportunities and a sustainable competitive advantage.

How it was last year

In 2025 we received 257 submissions, and 103 have advanced to the semifinal.

Our panel of 19 distinguished judges — including leaders from top industry organizations — selected the top three nominees in each of four categories.

AI Discovery Award 2025

Judging committee by Awards categories

To join the judging committee, please reach out to dpol@nebius.com.

BioPharma

Ivan Burkov

Inkef

Partner at Inkef, investing in early-stage AI-driven TechBio and HealthTech companies across Europe. With a background spanning research, pharma, VC, and startup building, Ivan partners closely with founders applying AI to drug discovery and healthcare delivery, bringing deep scientific expertise and extensive networks across biotech and pharma.

Athina Hadjixenofontos

AbbVie

Head of AI/ML in Biotherapeutics & Genetic Medicine discovery at AbbVie, driving data and model infrastructure to generate impact on patient lives. Previously at Recursion Pharmaceuticals.

Lev Kogon, PhD

Nykode Therapeutics

PhD in Health Economics. Biotech and pharma leader with 15+ years of global experience in strategy, business development, and licensing. VP at Nykode Therapeutics. Worked for Bayer, Daiichi Sankyo and GSK.

Anna Kostikova, PhD

Monte Rosa Therapeutics

PhD in computational biology and bioinformatics. Biotech and pharma leader with 15+ years of experience in AI-driven drug discovery and development. Senior Director at Monte Rosa Therapeutics. Co-founder and CSO at Asylia Diagnostics and InsideDNA. Worked for Novartis.

Pavel Sinitcyn, PhD

Utrecht University

Assistant Professor in AI Technology for Life, and Biomolecular Mass Spectrometry groups at Utrecht University.

Nick Venanzi

NVIDIA

Senior Solutions Architect at NVIDIA, designing AI and high performance computing solutions for computational biology and drug discovery. Previously a founding AI/ML engineer at a biotech startup, seasoned in building machine learning pipelines for therapeutics and shaping companies’ AI strategies.

Genomics

Evan Floden, PhD

Seqera

CEO & Co-Founder of Seqera and Nextflow. Previous experiences in biotech and medical devices. He is passionate about the next generation of AI-native life science research.

Ekaterina Kotelnikova

Clarivate

Senior consulting leader focused on AI-powered drug discovery and translational science. With 20+ years of experience, she brings deep expertise in knowledge graphs, biomarkers, and data-driven decision-making in pharma and biotech.

Javier Nunez-Vicandi

Sofinnova Partners

Principal at Sofinnova Partners, focused on digital medicine and investments at the intersection of healthcare and technology. Former VC at Heal Capital and M Ventures, and former clinical doctor at Oxford University Hospitals.

Digital Health

Kelly Kung

Accel

Investor at Accel, focused on enterprise software and AI infrastructure. Previously at General Atlantic and J.P. Morgan, Kelly also worked in AI research at Harvard and held commercial roles at early-stage startups in the US and UK.

Guy Spigelman

Ingenix AI

Chief Business Officer at Ingenix AI, a TechBio company developing AI-driven technologies to de-risk clinical trials. Previously at AWS, he worked on AI and cloud adoption in healthcare and life sciences, and co-founded AION Labs, an AI-focused biopharma venture studio.

Cedric Steenbeke

NVIDIA

A healthcare tech expert with 15+ years of experience, leads NVIDIA’s Healthcare & Life Sciences startup ecosystem in EMEA. Previously at Microsoft, he shaped healthcare data platforms and AI adoption, driving digital health innovation.

Emil Schaefer

Index Ventures

Emil is a Partner at Index Ventures in London. He works with founders primarily within Enterprise AI, Healthcare and Financial Services across Europe and the US.

Evgeny Shvarov

InterSystems

Senior Manager of InterSystems Corporate Venture Fund and co-founder of Caelestinus Digital Health Startup Incubator with over 10 years’ experience in digital health. Evgeny specializes in VC investment in early stage AI-driven and digital health interoperability focused solutions.

Madeleine Thun

AstraZeneca

Senior healthcare innovation leader with experience across AI, digital health, pharma, and global innovation ecosystems, focused on accelerating the adoption of emerging technologies in healthcare. Currently leading AstraZeneca’s European innovation network.

Medical Devices

Mikael Brudfors

NVIDIA

Solutions Architect at NVIDIA, working with medtech companies and clinical partners across Europe on AI-powered medical devices, real-time surgical systems, and accelerated computing for healthcare applications.

Uri Weinberg

Novocure

MD, PhD is Chief Medical Officer and Chief Innovation Officer at Novocure, where he oversees clinical development, medical affairs, and early-stage innovation for TTFields therapy. He joined the company in 2008 and holds an M.D. Ph.D. from the Technion — Israel Institute of Technology.

Medical Imaging

Dr. Ilya Burkov

Nebius

Global Head of Life Sciences and Healthcare Growth, with 15+ years of experience across clinical medicine, biotech and cloud. Previously at Cambridge University, Addenbrooke’s and AWS, he brings a unique blend of expertise to healthcare innovation. He is passionate about using AI and cloud technologies to drive scientific progress and improve patient outcomes.

Javier Fernandez-Marques

Flower Labs

Lead Research Scientist, working on the core framework and Flower Simulation Runtime. Focuses on the intersection of machine learning and systems, particularly efficient on-device ML workloads for federated learning. PhD in Computer Science from the University of Oxford. Javier was a Research Scientist at Samsung AI.

Ralph Schiess

Proteomedix

Founding CEO of ZH3D at the University of Zurich and former CEO and co-founder of Proteomedix, a Swiss diagnostics company acquired in 2023. He has more than 20 years of experience translating academic research into commercial healthcare products.

Paul Pouyanne

Cathay Innovation

Vice President at Cathay Innovation, focusing on frontier technologies and digital health investments. He has led investments in companies including Nabla, Bioptimus, Entalpic, and Tucuvi, supporting the growth of AI-driven healthcare and deep tech startups.

How Healthcare & Life Sciences leaders use Nebius solutions

See how HCLS leaders are overcoming compute challenges. These stories reveal the strategies shaping the future of the Healthcare & Lifesciences sector.

View customer stories

Nebius partners for Healthcare & Life Sciences

Our solutions for the Healthcare & Life Sciences industry go beyond products. Our partners are here to assist your organization at every level to build and execute transformative AI strategies, products and services.

Explore our partners

FAQ

The application deadline is on May 8, 2026.

Start your AI journey today

Talk to an expertTry console

Learn more

Terms and conditions​