Modular Raises $250M to Simplify AI Deployment Across Hardware

Modular secured $250 million in its third financing round for its AI deployment platform, led by Thomas Tull’s US Innovative Technology fund and DFJ Growth, valuing the company at $1.6 billion. Existing investors, including Google Ventures, General Catalyst, and Greylock Ventures, also participated.

Founded in 2022, Modular provides an AI deployment platform enabling developers to run AI applications across diverse computer chips. It includes CPUs, GPUs, ASICs, and custom silicon, eliminating the need for code rewrites. The company built a software infrastructure layer and a specialized programming language for enterprise AI model deployment.

Advancing Unified AI Deployment

Modular’s platform offers an enterprise-grade AI inference stack, abstracting hardware complexities. It addresses the fragmented ecosystem requiring specialized code for each architecture. This approach aims to loosen vendor lock-in. The platform supports architectures from Nvidia, AMD, and Apple’s custom silicon, solidifying its position as a versatile AI deployment platform.

The company reports performance gains of 20% to 50% over leading frameworks like vLLM and SGLang. These gains occur on next-generation accelerators, including Nvidia B200 and AMD MI355. Modular’s vision resonates with industry leaders. AMD, Nvidia, and Amazon.com Inc. joined as ecosystem partners.

Modular also collaborates with AI application developers such as Inworld AI for speech synthesis. San Francisco Compute Co., operating a GPU cluster marketplace, is another partner.

The company plans to expand hiring across North America and Europe.

It will scale its cloud platform and extend support for edge hardware. Modular will also broaden its focus from inference into AI training.

The funding brings Modular's total capital raised to $380 million.