Wafer Raises $40M Series A to Optimize AI Inference
Fundraise Insider tracks newly funded startups each week and delivers verified sales leads of C-suite contacts straight to your inbox, so you reach the right people while the funding is still fresh. See pricing.
Wafer has closed a $40 million Series A co-led by Marathon Management Partners and Chemistry. Wing Venture Capital, AMD Ventures, Outset Capital, Fifty Years and Y Combinator also took part, along with existing backers.
The company operates from San Francisco, and it sells technology that automates the work of optimizing how AI models are served in production.
That work is still mostly done by hand. Wafer says inference optimization today remains manual and service-intensive, and it is usually treated as a one-time exercise carried out before a model goes live. The company wants the opposite: optimization that never stops.
Its platform reads an application’s workload traffic patterns and performance constraints, then searches for the best deployment configuration across models, inference engines, kernels and hardware. Rather than tying itself to a single model, accelerator or inference engine, Wafer positions the product as a layer that sits across the stack.
The problem it is chasing has a clear financial shape. As more AI applications reach production, inference costs turn into a meaningful line item, particularly for high-volume products handling large numbers of user requests. Wafer’s pitch is performance per dollar, found continuously by software instead of by engineers tuning workloads one at a time.
The California company will put the new money toward automating more of that optimization loop, with the aim of giving every deployment the equivalent of an expert inference-performance team always hunting for improvements. AMD Ventures adds a strategic semiconductor investor to the register as Wafer builds out technology meant to work across different AI hardware environments.
A long list of technology founders and executives joined as angels, 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.
Teams that sell into software companies can pull verified contacts at newly funded infrastructure startups from B2B SaaS Leads.