Modal Labs, Inc.
High-performance, serverless AI infrastructure that allows developers to run inference, training, and batch processing with sub-second cold starts and instant GPU autoscaling.
At a Glance
- AI/ML Startups
- Scientific Research Institutions
- Media and Entertainment Companies
- Enterprise AI teams
AI Tools by Modal Labs, Inc.
(1)
Modal
Serverless Cloud for AI Workloads
Discussions
No discussions yet
Be the first to start a discussion about Modal Labs, Inc.
Latest News
Butter is joining Modal - AI sandbox technology acquisition
Product Updates: RTX Pro 6000 Blackwell support and Sandbox FS API
Runway Chooses Modal to Power Real-Time Inference for Runway Characters
Modal Labs targets $2.5 Billion valuation for AI inference infrastructure work
Products & Services
Scalable, low-latency inference for LLMs, audio, and image generation.
Platform for fine-tuning open-source models on single- or multi-node GPU clusters.
Ephemeral, secure execution environments for running untrusted code.
High-throughput on-demand batch workloads for large-scale data processing.
Market Position
Positions as the developer-first alternative to standard cloud providers (AWS/GCP/Azure) and Kubernetes-heavy setups, focusing on speed (100x faster startup than Docker) and ease of use (pure Python functions).
Leadership
Founders
Erik Bernhardsson
CEO of Modal. Previously CTO at Better.com and Engineering Manager at Spotify, where he led the music recommendation team and created Luigi and Annoy.
Akshat Bubna
Co-founder and CTO of Modal. MIT graduate (Bachelor and Master in CS). Joined as co-founder in August 2021.
Executive Team
Erik Bernhardsson
CEO
Ex-CTO Better.com, Ex-Spotify Engineering Manager.
Akshat Bubna
CTO
MIT CS graduate, key architect of Modal's infrastructure.
Board of Directors
Founding Story
Started by Erik Bernhardsson after his experience at Spotify and Better.com, where he realized existing cloud infrastructure was too slow and complex for data/AI teams. The vision was to create a platform where the cloud feels as fast and interactive as a local machine, removing the need for Kubernetes or complex YAML configurations.
Business Model
Revenue Model
Pay-per-use serverless compute model based on resource consumption (CPU, GPU, Memory, Disk) plus fixed monthly subscription tiers for advanced features.
Pricing Tiers
$30/mo free compute credits, 3 workspace seats, 100 containers, 10 GPU concurrency.
$100/mo free compute credits, unlimited seats, 1000 containers, 50 GPU concurrency, custom domains.
Volume-based discounts, custom GPU concurrency, HIPAA compliance, Okta SSO, dedicated Slack support.
Target Markets
- AI/ML Startups
- Scientific Research Institutions
- Media and Entertainment Companies
- Enterprise AI teams
- LLM Inference and Serving
- Model Fine-tuning (Whisper, Llama, etc.)
- Large-scale Batch Data Processing
- Secure Execution of Untrusted Code
- AI Agent Infrastructure
- Runway
- You.com
- Allen Institute for AI
- Lovable