CoreWeave at NVIDIA GTC 2026
What We Showcased and Why it Matters

From NVIDIA HGX B300 launches to live demos and expert-led sessions, see how CoreWeave brought AI Cloud innovation to life.

Watch the action

The shift from experimentation to production

As a Premier Partner at NVIDIA GTC 2026, CoreWeave showcased how an AI Cloud purpose-built for training, inference, and agents is powering the world’s most advanced models and applications.

Across launches, demos, and expert-led sessions, we highlighted how CoreWeave Cloud—The Essential Cloud for AI™— delivers the pace, performance, and partnership required to move from experimentation to production at scale.

From next-generation GPU deployments to real-world AI use cases, our presence at GTC reflected a broader industry imperative: AI is moving beyond experimentation, and the teams that win will be those that can build, run, and improve AI systems with speed, reliability, and control.

Read the full recap

New

Announced at GTC

NVIDIA HGX B300 on CoreWeave

CoreWeave unlocks NVIDIA HGX B300 performance for large-scale reasoning and agentic AI workloads, with expanded memory, faster networking, and production-grade reliability.

Built for the most demanding AI workloads, CoreWeave Cloud makes it easier for teams to train, deploy, and scale with unmatched efficiency on next-generation GPU infrastructure.

Read the announcement Watch the video

More launches shaping how AI is built and run

Environment-Free Reinforcement Learning

Train AI agents directly on production traces without building simulated environments, accelerating RL workflows while lowering operational complexity.

Learn more

Production Agent Evaluations with Weave

Analyze production traces, detect agent failure modes, and improve evaluation datasets to accelerate agent development and iteration.

Learn more

W&B Mobile App

Monitor training runs in real time from your phone with metrics, alerts, and experiment insights designed for always-on AI development.

Learn more

Product demos from the GTC floor

Explore demos that show how CoreWeave powers real-world AI workloads—from training and inference to agent-based systems and production-scale deployments.

Observability

How to Monitor AI Systems End-to-End, from Metal to Application

See how CoreWeave gives teams end-to-end visibility across infrastructure, workloads, models, and agents, making it easier to detect issues, diagnose performance gaps, and operate AI systems at scale.

Watch demo

AI model training

SUNK

How to Train AI Models at Scale Without Wasting Compute

08:20

INFERENCE

How to Run AI Inference in Production Without Rebuilding Your Stack

13:35

CoreWeave Cloud

How to Run AI Workloads at Scale: Inside the CoreWeave Cloud Platform

08:35

GENERATIVE AI

Media and ENTERTAINMENT

How to Train Custom Generative AI Models Without Managing Infrastructure

3:31

Dig into your use case with a CoreWeave expert

Where GTC came alive: Inside CoreWeave House

CoreWeave House pulsed at the center of GTC 2026—an always-on current of energy, ideas, and just enough chaos. We kicked off with a packed Jensen keynote watch party, the room buzzing in real time as the future unfolded on screen. Then we rolled into the Aston Martin Aramco F1 breakfast with speed, precision, and performance shaping every conversation. Our Full-Stack Developer Meetup brought together the pioneers pushing AI forward. And the wrapup party invited 1,000+ people in as Mix Master Mike took over and the whole place let loose. A week of AI momentum, distilled into one space.