Runway video models move into private deployment: enterprises license weights and run generation and editing on single-tenant GPUs

Enterprises can now deploy Runway's video generation and editing models inside their own isolated compute, rather than only calling them through Runway's hosted platforms. Runway previously offered these models through Runway Creative and Runway Dev, so enterprise use was limited to platform-side calls. The partnership covers Gen-4.5 and Aleph 2.0, explicitly including fine-tuned versions, in a single-tenant deployment form.

The deployment works by licensing model weights directly from Runway while Baseten supplies an isolated workload plane and dedicated GPUs, with Baseten managing the control plane and underlying infrastructure. This split separates weight licensing from serving infrastructure, and access to the weights is governed by a distribution agreement. Each deployment runs in an isolated workload plane with compute and scheduling isolated to the organization, on infrastructure covered by Baseten's SOC 2 Type II and HIPAA compliance programs.

Enterprises can fine-tune these models on their own data so outputs reflect their brand, canon, and content guidelines. The fine-tuning work is governed by the enterprise's direct license with Runway, while Baseten runs the fine-tuning and deploys the resulting model. Runway and Baseten support each deployment together, from model customization through production serving.

This news is for enterprise creative and engineering teams that need video generation and editing models inside their own isolated environment, in a single-tenant dedicated inference setting with fine-tuning on proprietary data. Aleph 2.0 supports multi-scene videos up to 30 seconds at 1080p. The source states no next step.

The source gives no comparison data on generation quality, latency, or throughput, and does not state the data volume or compute needed for fine-tuning. The uptime commitment is 99.99%, but its measurement basis and scope are not described.