Mako
Optimizes GPU kernels for faster, efficient AI inference.
Updated May 2026
Overview
- Website
- mako.dev
- Founded
- 2024
- Headquarters
- Manhattan, New York, United States
- Ownership
- Private
- Segment
- Tool Integration & Execution
Product overview
Mako is a startup developing AI-native infrastructure to automate GPU kernel generation and tuning, enabling peak performance across NVIDIA, AMD, and custom accelerators without code rewrites. Its flagship products, MakoGenerate and MakoOptimize, streamline deployment of PyTorch and Hugging Face models, reducing AI inference costs. Co-founded by Cornell Tech professor Mohamed Abdelfattah and others, it raised $8.5M in seed funding led by M13.
Revenue model
SaaS tools for GPU optimization and AI workload deployment
Moat
- Cost Advantages
- Switching Costs
- Proprietary Technology
- First Mover
- Scale Advantages
- Distribution
Mako refers to two entities with distinct moats: Mako Mining Corp.'s high-grade San Albino gold deposit provides a cost advantage through low AISC and operational resilience, while Stryker's Mako robotic surgery system builds a wide moat via high surgeon switching costs, a large installed base, proprietary implants, first-mover advantage, and global distribution network.
Headwinds
Extremely competitive market with GPU vendors like NVIDIA having strong optimization capabilities and ecosystem lock-in.