Wafer
AI that makes AI fast
Y CombinatorSummer 2025ActiveB2BEngineering, Product and DesignArtificial Intelligence
What they do
Problem
AI models often run inefficiently on diverse GPU hardware, causing slow and costly inference for chip makers and cloud providers.
Solution
Wafer provides AI agents that autonomously optimize GPU kernels for AI inference to maximize performance across hardware types.
Who it's for
Chip companies and cloud providers that need peak-performance AI inference on their hardware platforms.
Founders
Emilio Andere
Founder
prev @ argonne, uchicago sand lab, elicit. math at uchicago.
Steven Arellano
Founder
prev @ two sigma, google, sei labs, and axlab. cs + econ at uchicago.
Open roles 3
Company timeline
- Founded2025
- YC batchSummer 2025
- StatusActive
- Team size6 people
- LocationSan Francisco, CA, USA
Similar startups
Klavis AI
Live Environments for Training AI Agents
Cedana
Fast, reliable, reproducible AI with GPU live migration
Density AI
Density is developing general-purpose compute that purpose-built for…
Mem0
The Memory layer for AI Agents
Inferact
The AI industry has historically been bottlenecked by training.
Nozomio
product lab solving context in ai