Founded in 2022, Modular builds a software infrastructure layer and a specialized programming language that abstract hardware differences, allowing enterprise developers to deploy AI models on CPUs, GPUs, ASICs and custom silicon with no code migration. Its platform already supports Nvidia, AMD and Apple architectures and, in its latest release, claims 20–50% performance gains on next-generation accelerators such as Nvidia’s B200 and AMD’s MI355 when compared with frameworks like vLLM and SGLang. The company collaborates with ecosystem partners including AMD, Nvidia and Amazon, as well as AI developers like Inworld AI and GPU-cluster marketplace San Francisco Compute. Modular’s long-term goal is to reduce vendor lock-in and expand hardware choice for AI inference and, eventually, training workloads. The firm employs more than 130 people and is headquartered in the San Francisco Bay Area. With the new funding, it plans to scale hiring in North America and Europe, broaden cloud and edge hardware support, expand its cloud platform, and deepen its focus on AI training. To date, Modular has raised $380 million in venture capital, valuing the company at $1.6 billion after its Series B.

Total raised

$250M

Funding rounds

1

Latest round

Series B

Latest activity

Sep 2025