Modal Labs, Inc.

Modal Labs, Inc., doing business as Modal, is an American cloud computing company that operates a serverless platform for building, deploying and scaling compute-intensive applications, with a focus on AI and machine learning. Developers import Modal's Python SDK, annotate functions with decorators such as @app.function() and call .remote() to execute the code in the cloud; the platform builds containers from the local environment, schedules them onto specified CPU or GPU hardware and streams logs back in real time. Infrastructure as code is defined directly in Python rather than YAML, and the platform can scale from zero to thousands of containers or hundreds of GPUs.

The product suite spans Inference (serving LLMs and audio, image and video generation models), Training and fine-tuning (including multi-node clusters with RDMA networking and reinforcement-learning loops), Sandboxes (isolated ephemeral environments for untrusted code and AI agents), Batch processing (transcription, protein folding, weather forecasting), browser-based Notebooks, web endpoints exposing functions as HTTPS APIs, and storage through Volumes and Buckets. A Python SDK is the primary interface, with JavaScript/TypeScript and Go SDKs in beta as of October 2025.

The company is headquartered in New York City at 233 Spring St, with additional offices in Stockholm and San Francisco, and lists open engineering, go-to-market and G&A roles across those cities plus London.

Founding story

Modal was founded in January 2021 by Erik Bernhardsson, a Swedish engineer with a Master's in Physics from KTH Royal Institute of Technology who spent about seven years at Spotify leading data, A/B testing and music recommendation teams before joining Better.com as CTO in 2015, where he scaled engineering from one to over 300 people. Akshat Bubna joined as co-founder and CTO in August 2021; he holds an undergraduate degree in mathematics and computer science from MIT, was an early engineer at Scale AI, and has a competitive programming background including International Olympiad in Informatics medals. The company grew out of Bernhardsson's frustrations with traditional cloud infrastructure — YAML-heavy configuration management and slow scaling — and aimed to make cloud execution feel like local development.

Business model

Modal sells developer-first cloud infrastructure on a consumption basis: customers are billed by the second for actual compute used, with no long-term commitments or charges for idle resources. Revenue expands as customers run more functions, provision larger GPU workloads and adopt adjacent products such as Sandboxes, web endpoints, Volumes, Buckets, Notebooks and clustered computing. A multi-cloud architecture lets Modal source capacity across providers, and marketplace integrations with major clouds allow enterprises to apply existing committed spend.

Consumption-based pricing: customers pay per second of compute actually consumed across CPU and GPU resources, so monetization scales with workload volume and duration rather than seats or subscriptions. The website advertises $30/month of free compute for new users.

Traction

Reported metrics include approximately 153 employees as of April 2026, annualized revenue rising from about $60M in September 2025 (about $119M at end-2025 per Sacra's estimate) to roughly $300M by April/May 2026, more than one billion sandboxes launched, and capacity across 13 cloud providers. The company states it serves thousands of ML teams. One source cites Lovable creating 250,000 apps in a weekend on more than one million Modal sandboxes.

Latest developments

Third-party profiles report that revenue grew roughly fivefold from about $60M annualized in September 2025 to roughly $300M by April–May 2026, that the platform expanded to 13 cloud providers, and that cumulative sandbox launches passed one billion with Sandboxes contributing about one third of revenue. A $355M Series C at a $4.65B post-money valuation, led by General Catalyst and Redpoint Ventures, was reported in May 2026, following an initial tranche at $2.5B in February 2026. Recent product releases include Volumes v2, JavaScript/TypeScript and Go SDKs in beta, browser-based Notebooks with GPU memory snapshots, and clustered multi-node RDMA computing.