Dr. Ilya Burkov (Nebius): Powering Next‑Gen Biotech Research with High‑Performance AI and Cloud GPUs

Corporate Overview & Strategic Positioning

• Nebius is a Dutch-headquartered, NASDAQ-listed cloud provider (one of only two neocloud-like companies listed) operating at the intersection of hyperscale infrastructure and neocloud flexibility.
• The company is vertically integrated, managing its own data centers, racks, and servers in-house to ensure reliability and resilience.
• Nebius distinguishes itself from hyperscalers by avoiding vendor lock-in and offering flexibility without long-term contracts, while differentiating from neoclouds by providing a full-stack managed approach (Managed Kubernetes, Managed Slurm).
• Infrastructure capabilities include backwards compatibility with hyperscaler standards (e.g., S3 bucket support) and support for global expansion across the EU, US, Asia, and the Middle East.
• The company currently ranks among the top 40 supercomputers globally in two independent rankings.

Pricing, Flexibility, & Operational Model

• Nebius employs a "pay-as-you-go" or "Uber-style" model allowing customers to scale up for high-intensity workloads and scale down when capacity is not needed, without purchasing hardware.
• The platform supports "burst capacity" specifically designed for biotech and healthcare workflows that require temporary spikes in GPU usage followed by data analysis phases.
• Deployment is designed for non-experts, requiring no complex certifications to establish a cluster, allowing scientists to focus on their domain expertise rather than AI infrastructure.
• Customers can utilize reserve capacity for lower prices or on-demand access for the latest NVIDIA GPUs (e.g., H100s) with immediate availability for single nodes.

Healthcare, Life Science, & AI Applications

• Radiology & Diagnostics: AI is positioned as a tool to reduce radiologist fatigue from repetitive tasks (e.g., spotting fractures in high-volume scans), aiming to reduce errors rather than replace human professionals.
• Agentic AI (2025 Trend): Adoption is driven by the need to automate repetitive processes to alleviate workforce shortages in healthcare systems across Europe and Asia.
• Drug Discovery: AI integration aims to exponentially shorten time-to-market for drugs while reducing costs and accelerating the targeting of specific treatments.

Selected Case Studies & Performance Metrics

• Converge Bio (Israel): Deployed on Nebius infrastructure to create a single-cell model for oncology; the model outperformed the Zuckerberg and Chen single-cell GPT in accuracy and prediction using H100s over a two-week training round.
  • Capabilities include virtual screening to predict treatment efficacy before chemotherapy, reducing patient trial-and-error and hospital burden.
• Simulacra AI (London): Utilizes Nebius for quantum chemistry research to accelerate drug development and material design.
  • Reports 90% faster performance compared to previous experiences with hyperscalers.
  • Currently working on models with over 100 million parameters and plans to process massive datasets.
• Quantory (Yerevan, Armenia): Focuses on molecular generation and 3D shape design for drug discovery, leveraging large existing datasets.
  • Targets internal R&D teams within the pharma industry to reduce manual workload.
• Prima Mente (London): Developing a 7-billion-parameter model to train on large-scale wet lab data for neurodegenerative disease progression.
  • Collects data from Edinburgh hospitals and other European/US institutions.
• Genesis Therapeutics (US/Bay Area): Uses Nebius specifically for burst capacity that hyperscalers cannot provide flexibly enough.
  • Engages in AI-powered drug discovery while collaborating with major pharma companies.

Forward-Looking Statements & Market Trends

• Agentic AI is identified as a top strategic pick for 2025 in the healthcare sector.
• The company projects continued expansion of its AI footprint globally, specifically tailoring infrastructure for the unique scaling needs of life sciences.
• Future development focuses on handling "huge data sets" and "multi-omics" to further accelerate drug discovery and genomics.
• The company emphasizes a shift toward multilingual and non-English-specific AI applications, exemplified by Quantory's work with Armenian language models.