Modal

Modal is a serverless cloud for AI and data workloads that lets engineers run Python at scale on pooled CPUs and GPUs with sub-second container starts and per-second billing. Founded in 2021 by Erik Bernhardsson — former CTO of Better.com and creator of Spotify's recommendation system and the Annoy library — and Akshat Bubna, Modal exposes compute as a Python decorator: wrap a function, deploy it, and Modal handles containerization, scheduling, autoscaling, and GPU provisioning. The company reported surpassing $300 million in annualized revenue in 2026 and raised a $355 million Series C in May 2026 at a $4.65 billion post-money valuation led by General Catalyst and Redpoint Ventures, with notable revenue coming from sandboxes that execute AI-generated code.

Modal has become the default compute substrate for AI-native teams that don't want to manage Kubernetes or GPU reservations — a fast-growing alternative for the next generation of inference, fine-tuning, and agent workloads.

modal.com AI / GPU Cloud

United Statesprivate

At a glance

Total raised

$466M

last round series c

Headcount

51-200

approximate · public range

Headquarters

New York

United States

Founded

2021

5 years active

Status

Private

AI / GPU Cloud

Latest valuation

For subscribers

Subscribe to see the full profile

Including Funding rounds, Bull / Bear thesis, Stock + earnings, Roster changes, Patents, News, and Open roles.

Already subscribed? Sign in →

AI bull / bear

Bull case

Modal owns the most defensible position in serverless AI compute, and the market just repriced it accordingly: a $355M Series C in May 2026 led by General Catalyst and Redpoint at a $4.65B post-money — raised in two tranches beginning at $2.5B — against annualized revenue that quintupled from $60M in September 2025 to $300M. The engineering moat is real: sub-second container cold starts built on a custom file system and snapshotting stack make scale-to-zero GPU workloads actually viable, something generic clouds still have not matched, and per-second billing aligns Modal's revenue with the fastest-growing line item in software budgets. The second act is already material: Sandboxes — isolated execution for untrusted, model-generated code — has passed one billion launches and now contributes roughly a third of revenue, riding the steepest curve in the sector as every coding agent and agent platform needs exactly this primitive. Supply risk is managed by pooling capacity across thirteen cloud providers, so Modal arbitrages the GPU market rather than depending on it. Erik Bernhardsson's developer credibility from Spotify's recommendation systems and the Annoy library gives it authentic bottoms-up pull. If AI-native teams keep choosing a Python decorator over Kubernetes and GPU reservations, Modal becomes the default runtime for the post-AWS generation of AI products.

Bear case

Modal is squeezed from both ends of the stack. Above it, hyperscalers fold serverless GPU offerings into committed-spend agreements enterprises already hold; below it, GPU clouds rent raw capacity at prices a premium per-second layer cannot match for steady-state workloads. That exposes the graduation problem: serverless economics shine for bursty workloads, but once a customer's inference traffic stabilizes, reserved capacity gets cheaper and the largest spenders negotiate direct — Modal's best customers are structurally the most likely to leave, the Heroku pattern replayed for GPUs. The $4.65B valuation is roughly 15x an ARR figure that quintupled in eight months, which means the price assumes that trajectory continues; consumption revenue falls exactly as fast as it rises, with no seat-based floor, and Modal's customer base skews toward AI-native startups whose own funding is the most cyclical dollar in technology. Modal also carries the utilization risk of its pooled fleet across thirteen providers — idle GPUs are Modal's cost, not the customer's. The Sandboxes line, now a third of revenue, faces dedicated competitors like E2B, hyperscaler equivalents, and the realistic prospect that the largest agent platforms build isolation in-house once volume justifies it. Two-tranche rounds that step from $2.5B to $4.65B in months price in flawless execution on a commoditizing substrate.

Generated by editorial-sourced · July 1, 2026Refreshes quarterly · see disclosures