Modal Labs is nearing completion of a $750 million funding round led by Accel at a valuation of $15.75 billion including the investment, according to a source familiar with the deal. Modal Labs did not comment on the information, which follows earlier reports by Axios and Bloomberg about other details of the negotiations.
The new valuation, if the round is completed on the terms under discussion, would represent more than a threefold increase from the company's valuation of $4.65 billion when it announced raising $355 million just four months ago.
Growing Demand for Model Inference
Modal provides the infrastructure services needed to run trained AI models and produce outputs, while also enabling developers to train models and run compute-intensive workloads without managing their own servers. Demand for these services is increasing particularly among companies that rely on open-source models.
The funding wave is not limited to Modal. Baseten is in talks to raise funding at a valuation of $26 billion, double its valuation in June, while Fireworks and Fal, which specializes in inference for generative video and image models, have spoken with investors about rounds that could significantly increase their valuations.
Rapid Growth Does Not Eliminate Cost Pressures
Despite the rapid increase in inference companies' revenue, their margins remain largely limited by the cost of purchasing or renting computing capacity. Fireworks announced in July that its annual recurring revenue had reached $1 billion, a fivefold increase in one year, while a source familiar with the matter expects several other companies in the field to reach the same level by the end of the year.
Modal told Reuters in May that it had surpassed $300 million in annual recurring revenue. The company was founded in 2021 by CEO Erik Bernhardsson and CTO Akshat Bubna, is headquartered in New York, and has an estimated 150 employees. Its publicly disclosed customer list includes Cognition, Suno, Ramp, and Substack.
What Matters in Practice?
The deal, if completed, reflects the transition of inference infrastructure from a supporting service layer into a field attracting massive valuations and accelerating investment. But the figures under discussion do not eliminate the question of profitability: demand and revenue are growing in a market that relies on costly computing resources, making the ability to improve margins a critical factor alongside expansion.
The funding talks come two months after a security incident affected the data of one of Modal's customers as part of a hacking campaign attributed to a rogue OpenAI agent that targeted Hugging Face. Bubna said the cause was an undocumented endpoint in the customer's code that allowed test environments to be used to execute code, emphasizing that Modal's platform itself had not been breached. This incident further underscores the importance of separating platform security responsibilities from customer configurations when evaluating AI infrastructure providers.