Home / Companies / Vast.ai / Blog / February 2026

February 2026 Summaries

2 posts from Vast.ai

Filter
Month: Year:
Post Summaries Back to Blog
Vast.ai has introduced new updates aimed at enhancing orchestration support, refining billing and serverless workflows, and providing new templates and guides to streamline deployment processes. The company has launched the Vast.ai Startup Program, offering up to $2,500 in free GPU credits for early-stage teams moving their workloads to Vast.ai, along with 24/7 priority support and access to a wide range of GPUs. Enhancements to SkyPilot now allow users to filter capacity by Secure Cloud and provide full support for instance creation parameters, offering more control over deployment settings. The newly available Model Library features a model-first deployment approach, with templates organized by modality and use case, complete with recommended hardware presets. Vast.ai also continues to support AI workloads through its serverless API, facilitating automatic, pay-per-use scaling. Recent updates include fixes to the serverless GPU filter UI, billing transaction history improvements, and security enhancements.
Feb 13, 2026 614 words in the original blog post.
Vast.ai's Startup Program offers startups $2,500 in free GPU credits to move their AI workloads to its platform, providing a cost-effective alternative to major cloud GPU providers. Startups can access a wide range of GPUs, from consumer-grade to high-end accelerators, and benefit from 24/7 priority support during the credit period. The platform supports popular tools and workflows like Jupyter, Docker, and Hugging Face, allowing teams to train deep learning models and process large datasets efficiently while reducing infrastructure expenses by up to 81%. The program ensures security and compliance with SOC 2 Type II certification and offers additional protections for sensitive data, enabling startups to run workloads in secure, isolated environments. Vast.ai's serverless API facilitates seamless transitions from development to production by automatically routing inference workloads to cost-effective GPUs, eliminating manual management tasks. The program is designed to scale with startups, offering flexible, long-term infrastructure solutions without binding contracts, thus accelerating their growth and operational capabilities.
Feb 06, 2026 485 words in the original blog post.