# Modal Raises $750 Million at a $15.75 Billion Valuation

Published 2026-10-15 · AI.info Team · AI.info

> Modal announced a $750 million Series D led by Accel on October 14, valuing the AI infrastructure company at $15.75 billion. Its annualized run rate has passed $1 billion, while its Sandboxes product has reached a $300 million run rate, the company says.

[Modal](https://modal.com/blog/1-billion-annualized-run-rate-series-d)’s growth pitch rests on a tension: scarce GPUs make managed AI compute more valuable, but expensive compute can also squeeze the economics of providing it. On October 14, the company said it had raised $750 million in a Series D at a $15.75 billion post-money valuation, as it crossed $1 billion in annualized run rate. The company points to surging use of its platform; the cost of securing compute remains a pressure on providers.

## Accel leads a round less than five months after Modal’s last

Accel led the Series D, with Kleiner Perkins joining and all existing investors participating, Modal said. The investor group includes Menlo Ventures, General Catalyst, Redpoint, Lux Capital, Amplify Partners and Bain Capital Ventures. Modal will use the capital to build more infrastructure and products for AI applications.

The financing follows Modal’s May 21 Series C, when it raised $355 million at a $4.65 billion post-money valuation. The new valuation is more than three times that figure, in less than five months. Modal’s announcement does not give individual investment amounts or a detailed spending plan.

## Usage tripled as annualized run rate passed $1 billion, Modal says

Usage has tripled since its Series C and grown more than 15-fold year over year, the company says. “Since then, usage has tripled and we’ve grown more than 15x year-over-year,” Erik Bernhardsson, Modal’s co-founder and CEO, wrote in the announcement. Annualized run rate is a current pace projected across a year, not a report of revenue already earned over a full year.

Modal’s Sandboxes product has also reached a $300 million run rate on its own, the company says. Launched in 2023, Sandboxes provide isolated environments for running generated code and other workloads; the company says it has started close to 3 billion of them and can run up to 1 million concurrently. The product’s growth reflects a use for cloud compute beyond serving model responses: running code produced by AI agents.

Modal lists customers and workloads across several fields. DoorDash uses the platform for internal AI applications, while Cognition, Ramp, Linear and Hebbia build background agents on it. Runway uses Modal for video generation, and [Physical Intelligence](https://ai.info/companies/physical-intelligence), Generalist AI and 1X use it to train robots. TypeSafe AI’s Jev served 1 trillion tokens on the platform in its first three days, Modal also says.

## GPU shortages make the growth story more complicated

GPUs are scarce and models can be large and slow to load, Bernhardsson says. Modal’s answer, he wrote, is a software stack built around its own file system, container runtime, scheduler and networking, running across a capacity pool that spans more than 20 clouds. That approach lets the company draw on multiple providers rather than depend on a single cloud’s available capacity.

But access to compute does not remove its cost. Before Modal confirmed the round, [TechCrunch reported](https://techcrunch.com/2026/09/28/source-inference-provider-modal-labs-closing-in-on-750m-round-at-15-75b-valuation/) that high costs to acquire or lease compute were weighing on margins across inference providers. Modal’s funding announcement gives no margin figures, so its run-rate claim alone does not show how efficiently the company converts that demand into profit.

## New training and sandbox products extend Modal’s bet

Modal is also adding products beyond its existing inference and sandbox business. It recently launched multi-node training, which it says gives developers an RDMA-connected cluster without a capacity commitment, and is developing products for reinforcement learning and post-training. Modal servers and Endpoint Candidates add options for HTTP services and testing inference endpoints; virtual machines in Sandboxes are now generally available.

Modal’s team numbers more than 200 people across New York, San Francisco and Stockholm. Its next test is concrete: build out these products while meeting the compute demand behind the revenue pace it has announced. The Series D gives Modal more capital to do that, but the company has not disclosed how much of the $750 million it will put toward compute capacity.

## Sources

1. Original source · Blog post [Modal · Billion annualized run rate series d](https://modal.com/blog/1-billion-annualized-run-rate-series-d) modal.com

Other pages read

2. [Modal · Modal series c](https://modal.com/blog/modal-series-c) modal.com
3. [TechCrunch](https://techcrunch.com/2026/09/28/source-inference-provider-modal-labs-closing-in-on-750m-round-at-15-75b-valuation/) techcrunch.com

## Companies in this story

- [Modal Labs](https://ai.info/companies/modal-labs)

## Related stories

- [Broadcom Agrees to Lend Anthropic Up to $42 Billion for Chip Leases](https://ai.info/news/broadcom-anthropic-42-billion-compute-financing)
- [Epoch AI Estimates Huawei Horinger Has Just 7% of Colossus 2’s AI Compute](https://ai.info/news/epoch-ai-huawei-horinger-7-percent-colossus-2)
- [GMI Cloud Raises $668 Million for U.S. and Asia Expansion](https://ai.info/news/gmi-cloud-raises-668-million-global-ai-compute)
- [Crusoe Raises $3.9B at $30.9B Valuation for AI Factories](https://ai.info/news/crusoe-raises-3-9b-series-f-ai-factories)
- [Lambda Targets Up to $4 Billion Before a Planned 2027 IPO](https://ai.info/news/lambda-targets-4-billion-pre-ipo-raise)

## Explore

[More articles](https://ai.info/news)

---

Canonical: https://ai.info/news/modal-raises-750-million-series-d-15-75-billion-valuation
