Customer Cloud Solutions

7/30/2026

This post details the expansion of the Conductor platform to support LoRA training and inference for image generation, addressing challenges in AI asset creation for creative industries. It highlights secure storage of brand assets, seamless model customization, comparison of AI models, and inference for generating new images and videos. The new LoRA-centric version offers artist-level tools, pay-as-you-go pricing, and seamless cloud scaling. APIs will be available for developers to integrate Conductor training and inference into their workflows.

5/20/2026

This post announces a significant expansion of CoreWeave's AI infrastructure with a $6 billion investment in a new data center in Lancaster, Pennsylvania. The facility is purpose-built for AI workloads and is designed to be one of the largest AI data centers in the Mid-Atlantic region, with a capacity of up to 300 MW. This expansion aims to address the growing demand for compute power for AI development and training in the region, contributing to the establishment of a regional AI ecosystem.

5/20/2026

This post details how CoreWeave provides a full-stack compute solution for Jane Street, a quantitative trading firm. It highlights the delivery of scalable, secure, and high-performance cloud solutions, including direct connections, dedicated storage, and upgrade options. The solution enabled Jane Street to obtain timely access to cutting-edge infrastructure technologies for designing, executing, and scaling proprietary models, leading to improved compute performance and scalability for their training and inference workloads.

5/20/2026

Introduced Automated User Provisioning (AUP) and SUNK User Provisioning (SUP) to automate user provisioning for Slurm clusters. AUP integrates with enterprise identity providers (e.g., Okta, Microsoft Entra) via SCIM to synchronize users and groups with CoreWeave IAM. SUP extends this automation to Slurm-on-Kubernetes (SUNK) clusters, automatically creating POSIX users, groups, SSH keys, and Slurm users/accounts. This leverages NSSCache for resilient, local caching of identity data, improving performance and reliability compared to traditional daemon-based solutions like SSSD.

5/20/2026

This post introduces Kueue as a Kubernetes-native system for AI training workloads on Coreweave Kubernetes Service (CKS). It explains the challenges of traditional Kubernetes for GPU-intensive AI jobs (inelastic GPU nodes, 'all-or-nothing' requirements) and how Kueue addresses these with 'all or nothing' scheduling, ResourceFlavors for fine-grained resource management, and ClusterQueues/LocalQueues for multi-tenant resource sharing, including quota borrowing and preemption. Coreweave provides a Helm chart for Kueue installation and integrates its metrics into Grafana.

5/20/2026

This post introduces official support for SkyPilot, an open-source framework for multi-cloud AI orchestration, on CoreWeave's GPU infrastructure. Key technical contributions include: one-line InfiniBand enablement via a `<network_tier: best>` directive in SkyPilot resources, native integration with CoreWeave AI Object Storage using a `cw://` scheme for high-throughput data access (up to 7 GB/s per GPU), and autoscaling support for CoreWeave clusters. The post also details how SkyPilot abstracts cluster management, auto-selects resources, and handles job submission, with specific examples for training, development, and serving workloads, and highlights the benefits of combining SkyPilot's orchestration with CoreWeave's bare-metal Kubernetes architecture for price/performance optimization and reduced time-to-market.

5/20/2026

Introduces CoreWeave Interconnect, a private high-performance fiber-based network interconnect directly connecting CoreWeave Cloud to Google Cloud, leveraging Google Cloud's Partner Cross-Cloud Interconnect. Details SUNK Anywhere, extending the SUNK (Slurm on Kubernetes) system to support GPU environments across clouds or on-premises, enabling consistent cross-cloud AI training. Mentions LOTA Cross-Cloud for storage integration.

5/20/2026

This post details the architectural mismatch between general-purpose cloud platforms and AI workloads, explaining how limitations in provisioning, infrastructure rigidity, and cost control inherent in legacy clouds throttle AI innovation. It advocates for purpose-built AI clouds that offer on-demand GPU access, high-throughput networking, flexible resource orchestration, and deep observability to accelerate AI development and deployment.

10/15/2025

This post articulates the strategic rationale for CoreWeave's purpose-built AI cloud, contrasting it with general-purpose clouds. It details the CoreWeave Cloud platform's components: high-bandwidth networking (InfiniBand, RoCE, proprietary interconnects), latest GPUs, AI Object Storage with LOTA for high throughput, and SUNK (SLURM on Kubernetes) for unified orchestration. It also highlights integrations with Weights & Biases and OpenPipe for agentic workloads and AIOps, and the role of Mission Control for observability and orchestration.