Modal was founded by Erik Bernhardsson, the creator of the open-source ML tool Luigi at Spotify and former head of engineering at Better.com. Modal reimagines cloud computing for AI workloads, offering a serverless platform where developers can run Python functions on GPUs without managing containers, clusters, or infrastructure.
Core Products
Modal Cloud provides serverless execution of Python functions with automatic GPU provisioning. Developers write code locally and deploy with a single command, getting on-demand access to A100s, H100s, and other accelerators. Modal Volumes provide fast distributed storage for model weights and datasets. The platform handles scaling, scheduling, and resource management automatically.
Competitive Position
Modal’s developer experience is its primary differentiator. While other GPU clouds require infrastructure management, Modal abstracts this entirely, making it feel like running code locally but with cloud resources. This appeals strongly to ML engineers who want to focus on models rather than DevOps. The serverless model also means customers pay only for compute actually used.
Recent Developments
Modal raised its Series B in 2024, reaching unicorn status. The platform has gained strong adoption among AI startups and ML teams at larger companies. Modal expanded GPU availability, added persistent storage options, and improved cold-start times. The company has become particularly popular for batch inference, fine-tuning, and data processing workloads.
Outlook
Modal’s serverless approach to GPU computing could capture significant market share as AI development becomes more mainstream. The platform is particularly well-suited to the emerging pattern of running many short GPU jobs (fine-tuning, batch inference, preprocessing) rather than maintaining persistent clusters. Growth depends on expanding beyond early adopters to enterprise customers.