Modular
@modular.com
The unified AI inference stack - from custom GPU kernels to production cloud serving on NVIDIA and AMD. 2x performance. Top open models. Open source stack.
Modular's Company Logos
Modular's Brand Colors
Hex Code
Color name
RGB
HSL
CMYK
#637BFF
Cornflower Blue
99, 123, 255
231, 100, 69
61, 52, 0, 0
#FFFFFF
White
255, 255, 255
0, 0, 100
0, 0, 0, 0
#1032CF
Persian Blue
16, 50, 207
229, 86, 44
92, 76, 0, 19
About Modular
Modular is an AI infrastructure company building a unified inference stack that spans GPU kernels, model serving, and cloud deployment. Its platform is designed to run AI workloads across diverse hardware, including NVIDIA and AMD GPUs, Google TPUs, AWS Trainium, Qualcomm and Apple accelerators, and Intel, AMD, and ARM CPUs. Modular aims to help teams improve performance, portability, and cost efficiency while avoiding dependence on a single hardware vendor.
The company’s MAX platform supports model development and high-performance inference, with tools for serving more than 1,000 models and deploying custom or fine-tuned models. Its Mojo systems programming language and library of GPU kernels let developers optimize or create kernels for supported accelerators. Customers can deploy through Modular’s managed cloud, run inference in their own virtual private cloud, or self-host MAX and Mojo on their infrastructure.
Product options include shared endpoints, dedicated endpoints, and custom model deployments, with flexible usage-based or per-minute pricing.
Company type
Privately Held
Year founded
2022
Company size
51-200 employees
Related Brands
Tags
AI inference platform
Hardware-agnostic model serving
Custom model kernels
GPU/CPU optimization
Endpoint APIs
