
Mako is an AI company focused on solving the critical challenge of GPU code generation for modern compute. Their core offering consists of the Mako Kernel Agent, which uses LLM-driven code synthesis to create performant GPU kernels in various languages like CUDA, HIP, and Triton. This is integrated into the Mako Optimization Platform, which automatically benchmarks, tunes, and deploys these kernels into real workloads. The company aims to accelerate AI innovation by automating the generation of high-performance kernels, thereby removing a significant bottleneck in the field. They were formerly known as A2 Labs.

Mako is an AI company focused on solving the critical challenge of GPU code generation for modern compute. Their core offering consists of the Mako Kernel Agent, which uses LLM-driven code synthesis to create performant GPU kernels in various languages like CUDA, HIP, and Triton. This is integrated into the Mako Optimization Platform, which automatically benchmarks, tunes, and deploys these kernels into real workloads. The company aims to accelerate AI innovation by automating the generation of high-performance kernels, thereby removing a significant bottleneck in the field. They were formerly known as A2 Labs.
Core product: LLM-driven Mako Kernel Agent for generating GPU kernels
Platform: Mako Optimization Platform for benchmarking, tuning, and deploying kernels
Targets: Generates CUDA, HIP, Triton and supports multiple GPU vendors
Founded: 2024
Employee count: 28
Total funding: USD 8,500,000
GPU code generation and optimization for AI workloads
2024
Artificial intelligence / Developer tools
“M13, K5 Global, Flybridge, Neo, Irregular Expressions”