Lynx Analytics is a life sciences consultancy. Before the March 2026 business split, it ran LynxKite graph AI workloads on Nebius AI Cloud. Nebius’s customer story describes virtual machines, Managed Kubernetes, shared filesystems and Slurm support for model training and graph analytics. LynxKite 2000:MM let experts build and deploy predictive Graph AI models without writing code. LynxKite became an independent product company on March 1, 2026, so the original story records the earlier combined business.
Nebius reports that LynxKite 2000:MM scaled training and inference workloads from one to over 1,000 GPUs while sustaining utilization above 80%. Managed Kubernetes automated container orchestration and model lifecycles. Shared filesystems provided a unified data layer across virtual machines, while Soperator supplied managed Slurm capabilities. While reproducing tumor-modeling research, Lynx Analytics expanded LynxKite’s toolbox with neural ODEs, heterogeneous graph convolutions and LSTMs. It also added custom model inputs for detailed batching and real-time graph creation for individual samples.
In March 2025, the team described multi-GPU matrix multiplication, research experiments and small-molecule generation on Nebius Cloud. Its described workflow passes knowledge-graph link predictions through foundation models and automated third-party evaluation, feeding results back into the prediction loop. European cluster availability supported regional requirements for research experimentation. It discussed AI Studio as a prospective model-serving option and planned reusable biomedical knowledge-graph workflows. GeekWire’s April 2025 summary describes Nebius AI Studio supporting biomedical knowledge graphs and generative molecule design. The March 2026 split announcement also named a strategic partnership with Nebius.
Nebius describes Lynx Analytics using AI Cloud, Managed Kubernetes, shared filesystems and Soperator for LynxKite training and inference. The story also describes Lynx Analytics reproducing tumor-modeling research and adding neural ODEs, heterogeneous graph convolutions, LSTMs, custom batching inputs and real-time graph creation to LynxKite.
Nebius Customer story7 Dec 2025 · Nebius · Web pageNebius describes Lynx Analytics using AI Cloud, Managed Kubernetes, shared filesystems and Soperator for LynxKite training and inference. The story also describes Lynx Analytics reproducing tumor-modeling research and adding neural ODEs, heterogeneous graph convolutions, LSTMs, custom batching inputs and real-time graph creation to LynxKite.
Nebius Lynx Analytics team talks about Nebius
9 Apr 2025 · Nebius · Video
5:31: team describes deploying an existing Docker image on a Nebius instance. At 7:50–8:17 it describes GPU computation with BioNeMo models and small-molecule generation using LynxKite. Proposed AI Studio serving and a future biomedical knowledge graph are separate plans, not completed use. The recording predates the later company split.
Nebius Lynx Analytics team talks about Nebius9 Apr 2025 · Nebius · Video · Starts at 5:315:31: team describes deploying an existing Docker image on a Nebius instance. At 7:50–8:17 it describes GPU computation with BioNeMo models and small-molecule generation using LynxKite. Proposed AI Studio serving and a future biomedical knowledge graph are separate plans, not completed use. The recording predates the later company split.
Cloud for Scientific AI and Healthcare
13 Oct 2026 · nebius.com · Web page
Nebius describes Lynx Analytics using its virtual machines and managed Kubernetes for LynxKite 2000:MM Graph AI workloads, with Slurm support through Soperator. The platform let experts build and deploy predictive Graph AI models without writing code. Nebius reports average GPU utilization above 80% and scaling from one to over 1,000 GPUs.
Cloud for Scientific AI and Healthcare13 Oct 2026 · nebius.com · Web pageNebius describes Lynx Analytics using its virtual machines and managed Kubernetes for LynxKite 2000:MM Graph AI workloads, with Slurm support through Soperator. The platform let experts build and deploy predictive Graph AI models without writing code. Nebius reports average GPU utilization above 80% and scaling from one to over 1,000 GPUs.
Nebius Spotlight GTC 2025: Lynx Analytics
28 Mar 2025 · GeekWire Studios · Video
Lynx Analytics’ team describes Nebius Cloud use for research and development, including knowledge-graph link predictions evaluated through foundation models and an automated feedback loop. Nebius’s European clusters support regional requirements for experimentation. Evidence at 03:23.
Nebius Spotlight GTC 2025: Lynx Analytics28 Mar 2025 · GeekWire Studios · Video · Starts at 03:23Lynx Analytics’ team describes Nebius Cloud use for research and development, including knowledge-graph link predictions evaluated through foundation models and an automated feedback loop. Nebius’s European clusters support regional requirements for experimentation. Evidence at 03:23.
Nebius spotlight interview summary
9 Apr 2025 · GeekWire · Web page
An event page discusses multi-GPU computation and potential AI Studio applications.
Lynx Analytics is an AI consulting and services company serving pharmaceutical, biotechnology and other life sciences organizations. It builds data pipelines, analytical models and software for healthcare professional engagement, drug launch planning, marketing budget allocation and clinical trial optimization. Its users include medical science liaison, brand, marketing and sales teams. LynxScribe supplies reusable software components and data connectors for client projects. Since March 2026, Lynx Analytics has operated independently of LynxKite, its former graph analytics product business.
Historical LynxKite graph-AI workflows; now a services company · Health
How they use Nebius
Lynx Analytics is a life sciences consultancy. Before the March 2026 business split, it ran LynxKite graph AI workloads on Nebius AI Cloud. Nebius’s customer story describes virtual machines, Managed Kubernetes, shared filesystems and Slurm support for model training and graph analytics. LynxKite 2000:MM let experts build and deploy predictive Graph AI models without writing code. LynxKite became an independent product company on March 1, 2026, so the original story records the earlier combined business.
Nebius reports that LynxKite 2000:MM scaled training and inference workloads from one to over 1,000 GPUs while sustaining utilization above 80%. Managed Kubernetes automated container orchestration and model lifecycles. Shared filesystems provided a unified data layer across virtual machines, while Soperator supplied managed Slurm capabilities. While reproducing tumor-modeling research, Lynx Analytics expanded LynxKite’s toolbox with neural ODEs, heterogeneous graph convolutions and LSTMs. It also added custom model inputs for detailed batching and real-time graph creation for individual samples.
In March 2025, the team described multi-GPU matrix multiplication, research experiments and small-molecule generation on Nebius Cloud. Its described workflow passes knowledge-graph link predictions through foundation models and automated third-party evaluation, feeding results back into the prediction loop. European cluster availability supported regional requirements for research experimentation. It discussed AI Studio as a prospective model-serving option and planned reusable biomedical knowledge-graph workflows. GeekWire’s April 2025 summary describes Nebius AI Studio supporting biomedical knowledge graphs and generative molecule design. The March 2026 split announcement also named a strategic partnership with Nebius.
Nebius describes Lynx Analytics using AI Cloud, Managed Kubernetes, shared filesystems and Soperator for LynxKite training and inference. The story also describes Lynx Analytics reproducing tumor-modeling research and adding neural ODEs, heterogeneous graph convolutions, LSTMs, custom batching inputs and real-time graph creation to LynxKite.
Nebius Customer story7 Dec 2025 · Nebius · Web pageNebius describes Lynx Analytics using AI Cloud, Managed Kubernetes, shared filesystems and Soperator for LynxKite training and inference. The story also describes Lynx Analytics reproducing tumor-modeling research and adding neural ODEs, heterogeneous graph convolutions, LSTMs, custom batching inputs and real-time graph creation to LynxKite.
Nebius Lynx Analytics team talks about Nebius
9 Apr 2025 · Nebius · Video
5:31: team describes deploying an existing Docker image on a Nebius instance. At 7:50–8:17 it describes GPU computation with BioNeMo models and small-molecule generation using LynxKite. Proposed AI Studio serving and a future biomedical knowledge graph are separate plans, not completed use. The recording predates the later company split.
Nebius Lynx Analytics team talks about Nebius9 Apr 2025 · Nebius · Video · Starts at 5:315:31: team describes deploying an existing Docker image on a Nebius instance. At 7:50–8:17 it describes GPU computation with BioNeMo models and small-molecule generation using LynxKite. Proposed AI Studio serving and a future biomedical knowledge graph are separate plans, not completed use. The recording predates the later company split.
Cloud for Scientific AI and Healthcare
13 Oct 2026 · nebius.com · Web page
Nebius describes Lynx Analytics using its virtual machines and managed Kubernetes for LynxKite 2000:MM Graph AI workloads, with Slurm support through Soperator. The platform let experts build and deploy predictive Graph AI models without writing code. Nebius reports average GPU utilization above 80% and scaling from one to over 1,000 GPUs.
Cloud for Scientific AI and Healthcare13 Oct 2026 · nebius.com · Web pageNebius describes Lynx Analytics using its virtual machines and managed Kubernetes for LynxKite 2000:MM Graph AI workloads, with Slurm support through Soperator. The platform let experts build and deploy predictive Graph AI models without writing code. Nebius reports average GPU utilization above 80% and scaling from one to over 1,000 GPUs.
Nebius Spotlight GTC 2025: Lynx Analytics
28 Mar 2025 · GeekWire Studios · Video
Lynx Analytics’ team describes Nebius Cloud use for research and development, including knowledge-graph link predictions evaluated through foundation models and an automated feedback loop. Nebius’s European clusters support regional requirements for experimentation. Evidence at 03:23.
Nebius Spotlight GTC 2025: Lynx Analytics28 Mar 2025 · GeekWire Studios · Video · Starts at 03:23Lynx Analytics’ team describes Nebius Cloud use for research and development, including knowledge-graph link predictions evaluated through foundation models and an automated feedback loop. Nebius’s European clusters support regional requirements for experimentation. Evidence at 03:23.
Nebius spotlight interview summary
9 Apr 2025 · GeekWire · Web page
An event page discusses multi-GPU computation and potential AI Studio applications.
Lynx Analytics is an AI consulting and services company serving pharmaceutical, biotechnology and other life sciences organizations. It builds data pipelines, analytical models and software for healthcare professional engagement, drug launch planning, marketing budget allocation and clinical trial optimization. Its users include medical science liaison, brand, marketing and sales teams. LynxScribe supplies reusable software components and data connectors for client projects. Since March 2026, Lynx Analytics has operated independently of LynxKite, its former graph analytics product business.
