CoreWeave Fully Connected 2026: CoreWeave Forge Launches, Turning the AI Loop Production Run Into a Better Model and Agent

CoreWeave Inc., an essential cloud for AI, announced CoreWeave Forge, a development layer for teams building and improving models and agents.It runs the entire AI loop – run, observe, curate, improve, evaluate and repeat – in one connected environment, and it’s open across the models, frameworks, and other clouds teams use. MasterClass and Canva are already building on CoreWeave Forge. The news was shared during Fully Connected, CoreWeave’s AI cloud conference, which brings together more than 4,500 customers, partners, developers and AI leaders to share how they are building and running AI in production.

Model and agent development has been split across tools from separate vendors, none built to work together. For the researchers and engineers carrying a model or agent through its full lifecycle, that fragmentation has a cost: a production trace does not feed the next training run, a finished experiment does not inform the next evaluation, and every handoff loses a signal or costs more time.

CoreWeave Forge unifies Weights & Biases Models, post-training expertise from OpenPipe, and the open-source marimo notebook project with CoreWeave’s own services, in one connected environment built for continuous improvement. This development layer closes the AI loop and assists customers in improving their models or agents, no matter their expertise, while keeping every team free to build with any model, framework or cloud, so improvement compounds with every version. Training runs, experiment tracking, evaluations and agent traces live where the model or agent actually runs. CoreWeave Registry keeps every checkpoint and agent configuration versioned. Every piece of it stays open to the models, frameworks, and clouds a team already uses.

“As more people build with AI, they’re putting models to work with their own data and workflows. That’s where the gaps between model capability and system performance become clear,” said Chen Goldberg, EVP, product and engineering, CoreWeave. “Engineering teams need to understand those gaps, identify the signals that matter, improve the next version and measure whether the change worked under real operating conditions. All of this has to function as a single, connected system. Forge brings all these capabilities on one platform, so what a business learns from running AI becomes part of how engineers improve it.”

What ships with CoreWeave Forge
CoreWeave Forge is available with new and expanded capabilities across the AI loop:

• CoreWeave ARIA, now generally available. ARIA is a coding agent that automatically analyzes large-scale experiment data and agent observability data, surfaces what drove a change, and proposes the next experiments worth running and the right agent worth running, with the evidence to support it
• Weights & Biases Models. Weights & Biases Models helps teams reliably and efficiently track and compare tens of thousands of experiments and millions of metrics, accelerating model development with advanced visualizations and automated workflows
• CoreWeave Notebooks. Notebooks is a developer experience that lives in the environment the work runs in, so a prototype carries straight into training, evaluation and production instead of dying in a rewrite. Custom visualizations are built on marimo
• CoreWeave Agent Lens.A new intelligent observability and continuous improvement tool for production agents. It analyzes tens of millions of traces, automatically turning them into insights that drive proven fixes. Detecting 20% more critical failures, CoreWeave Agent Lens fixes issues at one-tenth the cost based on a benchmark vs. a general-purpose frontier LLM
• CoreWeave Sandboxes, now available. A fresh, isolated environment for every run, whether agent tool use, reinforcement learning (RL) or model evaluation, already connected to the rest of the loop, so nothing extra has to be stood up to use it
• CoreWeave Registry.Every model checkpoint and agent configuration is versioned in open, portable formats, with clear lineage, so the next attempt can start from any point in a team’s history, on any stack
• CoreWeave Post-Training.Multiple services deliver post-training capabilities, including serverless RL, serverless supervised fine-tuning and model distillation, so improvement can go as deep as the weights. Serverless RL trains 1.4 times faster at 40% lower cost than a self-managed setup
• CoreWeave Inference. Now part of CoreWeave Forge, providing playground and inference deployment access to the latest open-weight models. Dedicated Inference is also adding a new capability in preview: RL Rollouts, which hot-loads checkpoints into a live deployment so a training loop can keep running without redeploying
• CoreWeave Forge Free, Pro and Enterprise. Three editions, so teams can start without a procurement cycle and grow into the platform. For the first time, teams can grow or expand across the loop