Modal Labs pursues $2.5B valuation as AI inference infrastructure attracts investor interest

Exclusive: AI inference startup Modal Labs in talks to raise at $2.5B valuation, sources say

Modal Labs, a startup specializing in AI inference infrastructure, is in the process of raising a new round at a valuation of about $2.5 billion, according to four people with knowledge of the deal. Should the deal close at these terms, the funding round would more than double the company's valuation of $1.1 billion announced less than five months ago. General Catalyst is in talks to lead the round, the people told TechCrunch. Modal's annualized revenue run rate (ARR) is approximately $50 million, our sources said. The discussions are early, and terms could still change. Modal and General Catalyst did not respond to our requests for comment. Model is focused on optimizing inference, the process of running trained AI models to generate answers from user requests. Improving inference efficiency reduces compute costs and reduces the lag time between a user's prompt and the AI's response. Modal is one of the handful of inference-focused companies attracting intense investor attention now. Last week, its competitor Baseten announced a $300 million raise at a $5 billion valuation, more than doubling the $2.1 billion valuation it reached just months prior in September. Similarly, Fireworks AI, an inference cloud provider, secured $250 million at a $4 billion valuation in October. In January, the creators of the open-source inference project vLLM announced they had transitioned the tool into a VC-backed startup, Inferact, raising $150 million in seed funding led by Andreessen Horowitz at an $800 million valuation. Meanwhile, TechCrunch reported that the team behind SGLang has commercialized as RadixArk, which sources told us secured seed funding at a $400 million valuation led by Accel. Modal was co-founded by CEO Erik Bernhardsson in 2021 after he spent more than 15 years building and leading data teams at companies including Spotify and Better.com, where he was the CTO. The startup counts Lux Capital and Redpoint Ventures among its earlier backers.

Modal Labs Targets $2.5 Billion Valuation for AI Inference Work | PYMNTS.com

The company, which specializes in artificial intelligence (AI) inference infrastructure, is in talks with venture capitalists about a new funding round, TechCrunch reported Wednesday (Feb. 11), citing sources familiar with the matter. According to the report, the deal would more than double Modal Lab's previous valuation of $1.1 billion, reached months ago when the company announced an $87 million Series B round. The report added that Modal Labs Co-founder and CEO Erik Bernhardsson denied that the company was actively raising funds and said his recent interactions with venture capitalists were simply general conversations. As covered here late last year, inference refers to the stage where a trained model processes new data and generates results. Examples of inference at work include a customer service chatbot replying to a query, or an AI system analyzing a financial document. "While training creates the model by processing vast datasets to learn patterns, inference applies that learned knowledge to perform specific tasks at scale," PYMNTS wrote. "As enterprises deploy AI systems that manage thousands or millions of requests daily, inference becomes the dominant operational challenge and cost driver." TechCrunch notes that Modal is among a small group of inference-focused startups catching the eye of investors. Last month, rival firm Baseten announced it had raised $300 million -- half of it from Nvidia -- valuing the company at $5 billion. And in October, cloud inference company Fireworks AI raised $250 million to expand its platform, with that round valuing the startup at $4 billion. Fireworks helps organizations use and customize large language models more efficiently, reducing costs and delays in deploying them. Last year, PYMNTS looked at inference and why it is now more important than training for most enterprises. Training a large language model happens just on a periodic basis, while inference takes place continuously each time a user interacts with an artificial intelligence system. "A single model might manage millions of inference requests per month, each requiring computational resources, adding latency and incurring costs," that report said. "For companies running artificial intelligence in customer-facing applications, inference performance directly affects user experience, system reliability and operational expenses." Inference, the report added, is emerging as a competitive category all its own. Brookfield has forecast that by 2030 around three-quarters of AI compute demand will come from inference, moving the economics of artificial intelligence "from training breakthroughs to the efficiency of serving models at scale," PYMNTS wrote.