Modal

About Modal

Modal was founded in 2021 by the former CTO of Khanmigo's parent organisation and a team of engineers from Stripe and other high-growth companies. The platform provides a Python-native serverless cloud designed specifically for computational workloads—model training, inference, data processing, and automation—that are difficult to run efficiently on traditional cloud platforms. Modal's core model is that developers define their compute workloads as Python functions decorated with Modal primitives, and the platform handles containerisation, dependency management, auto-scaling, and infrastructure. Functions can be scheduled, triggered by web hooks, or called from other systems. GPU access (A10G, A100, H100) is available on-demand with per-second billing, making it cost-efficient for bursty workloads like inference serving or batch training jobs. Modal has become popular in the AI developer community for building inference APIs on top of open-source models, running fine-tuning jobs, creating data pipelines for ML training data preparation, and running AI agents. The developer experience is notably polished compared to traditional cloud infrastructure. For European businesses, Modal's US incorporation (Delaware) and US-only data centre infrastructure present GDPR compliance challenges. There are no EU data centre options, and processing personal data through Modal requires GDPR transfer mechanisms. Modal does not publish ISO 27001 or SOC 2 Type II certifications as of early 2026. The platform is best suited for AI developers building non-personal-data workloads, internal tooling, or experimental systems where EU residency is not a requirement.