AI Infrastructure Surge: Modal Labs in Talks for $2.5B Valuation

In a move that signals a tectonic shift in the artificial intelligence gold rush, Modal Labs, the "serverless" cloud startup, is reportedly in advanced discussions to raise fresh capital at a staggering $2.5 billion valuation.

Just five months ago, the New York-based firm was celebrating "unicorn" status with a $1.1 billion valuation. Now, according to sources familiar with the matter, that figure is set to more than double. The round is expected to be led by General Catalyst, a heavyweight in the venture world known for backing infrastructure giants like Stripe and Databricks.

Beyond the Hype: Solving the "Inference" Bottleneck

While the last two years were defined by a race to train massive AI models, 2026 has become the year of running them. This process, known as inference, is the quiet engine behind every ChatGPT response, every AI-generated image, and every automated code snippet.

For enterprises, however, inference is a logistical nightmare. It is expensive, requires specialized GPUs that are in short supply, and often suffers from "latency"—the frustrating delay between a user's question and an AI’s answer.

Modal Labs has positioned itself as the "new utility" for this era. Led by CEO Erik Bernhardsson, the former data mastermind behind Spotify’s recommendation engine, and CTO Akshat Bubna, Modal offers a "code-first" platform. It allows developers to turn any Python function into a cloud-ready application that can scale from zero to thousands of GPUs in seconds.

The Numbers Behind the Surge

The investor frenzy isn't just based on potential. Modal is currently operating at an annualized revenue run rate (ARR) of approximately $50 million, showing significant commercial traction among AI-native companies like Scale AI and Meta.

The startup’s "serverless" model is its secret weapon:

Pay-per-use: Companies only pay for the seconds a GPU is actually running, a stark contrast to traditional cloud providers where reserved instances often sit idle.

Speed: Modal claims "sub-second cold starts," meaning AI models can wake up and perform a task almost instantly.

Capacity: By pooling GPU resources from various providers (including Oracle Cloud and AWS), Modal ensures its customers never run out of "compute" power.

A Growing Arms Race

Modal isn't alone in this high-stakes arena. The move comes on the heels of several massive raises in the inference space:

Baseten recently secured $300 million at a $5 billion valuation.

Fireworks AI reached a $3.4 billion valuation last year.

Together AI currently sits at a $3.5 billion valuation.

As AI transitions from experimental toys to critical business infrastructure, the companies that control the "plumbing", the chips, the networking, and the orchestration, are becoming the most valuable players in the ecosystem.

While CEO Erik Bernhardsson has reportedly noted that the company is not "actively" raising, he acknowledged the frequent outreach from VCs. If the deal closes at the $2.5 billion mark, it will cement Modal Labs as one of the definitive pillars of the modern AI stack, transforming how the world builds and scales intelligent software.