Wafer has raised $40 million in Series A funding to accelerate development of its AI-powered inference optimization platform. The round was co-led by Marathon Management Partners and Chemistry, with participation from Wing Venture Capital, AMD Ventures, Outset Capital, Fifty Years and Y Combinator. Existing investors also participated in the financing.
The round also included a group of technology executives and founders as angel investors, including Jeff Dean of DiscoveryLoop, Guillermo Rauch of Vercel, Andy Fang of DoorDash, Kyle Vogt, Akshay Kothari of Notion, Matthew Prince of Cloudflare and Scott Stephenson of Deepgram.
Wafer is developing technology designed to automate the process of optimizing AI inference deployments.
The company said much of inference optimization today remains manual and service-intensive, with optimization frequently performed as a one-time process before a model is deployed.
Wafer’s approach is designed to use AI to continuously optimize AI infrastructure.
Its platform analyzes an application’s workload traffic patterns and performance constraints and searches for the optimal deployment configuration across models, inference engines, kernels and hardware.
The company is targeting one of the increasingly important challenges facing AI infrastructure operators: improving model performance while reducing the cost of inference.
As more AI applications move into production, inference costs can become a significant component of operating expenses, particularly for high-volume applications processing large numbers of user requests.
Wafer’s technology is intended to continuously evaluate deployment configurations and identify opportunities to increase performance per dollar rather than relying on engineers to manually optimize individual workloads.
The new capital will support Wafer’s efforts to automate more of the inference optimization loop, with the goal of giving each deployment the equivalent of an expert inference-performance team continuously searching for improvements.
The company is positioning its platform as an optimization layer that can work across different parts of the AI infrastructure stack instead of focusing solely on a particular model, accelerator or inference engine.
Participation from AMD Ventures also adds a strategic semiconductor investor to the round as Wafer expands technology designed to optimize workloads across AI hardware environments.
Wafer said the funding will help accelerate development of its platform and support its broader goal of making continuous AI inference optimization available at scale.
