Wafer has Raised $4 Million in Seed Funding Led by Fifty Years to Build AI that Optimizes AI Infrastructure
- Karan Bhatia
- Apr 16
- 2 min read
Wafer, helping you ship the fastest inference in the world, led by Steven Arellano and Emilio Andere, has raised $4 million in seed funding led by Fifty Years, with participation from Liquid2 and Y Combinator. The round also includes participation from prominent angel investors such as Jeff Dean (Chief Scientist at Google), Wojciech Zaremba (Co-founder of OpenAI), Arash Ferdowsi (Co-founder of Dropbox), Dan Fu (Head of Kernels at Together), Kawal Gandhi (Office of the CTO at Google), Alfredo Andere (Co-founder and CEO of Latch Bio), Mokshith Voodarla (Co-founder and CEO of Sieve), Max Buckley (Head of Knowledge Research at Exa), and Tarun Chitra (Founder and CEO of Gauntlet), among others.
This investment will accelerate the development of an AI performance engineering agent, enabling hardware providers, cloud platforms, and frontier labs to maximize system efficiency and narrow the gap between current AI performance and its physical limits.
Why Wafer Started
Wafer is built on the belief that the key constraint in AI is no longer intelligence itself, but efficiency, captured as “intelligence per watt.” Increasing this metric expands the range of solvable problems, yet a significant gap remains between current AI system performance and hardware potential.
That gap is constrained by a shortage of highly specialized performance engineers, whose work, profiling, optimization, and hardware-specific tuning, is complex, manual, and difficult to scale. As hardware evolves rapidly across chips and architectures, traditional approaches like compilers struggle to keep pace.
Wafer’s approach is to build AI systems that optimize AI infrastructure itself, starting with an agent that performs the role of a performance engineer across diverse hardware environments.
Founded by longtime collaborators with backgrounds in AI infrastructure, research, and large-scale systems, the company is focused on a single mission: maximizing intelligence per watt to drive more efficient and scalable AI systems.
