# Wafer

> Agent-tuned inference stacks that run open models faster on the same weights.

Canonical: https://www.subconscious.dev/providers/wafer · By The Subconscious Team · Updated October 11, 2026

- Founded: 2025
- Example models: GLM-5.3, DeepSeek V4.1 Flash
- Website: https://www.wafer.ai

## Overview

Wafer builds AI agents that act as GPU performance engineers, then sells inference on the stacks those agents tune. The company came out of Y Combinator's Summer 2025 batch as a "Cursor for CUDA" that turned slow PyTorch into custom kernels. Founders Emilio Andere and Steven Arellano are based in San Francisco. Wafer raised a $4M seed led by Fifty Years in April 2026 and a $40M Series A co-led by Marathon and Chemistry in September 2026 at a $220M post-money valuation, with AMD Ventures and Wing joining. Its agents profile a workload, generate candidate configs across batching, decoding, quantization, engines, kernels and hardware, measure each one and deploy the winner.

Wafer calls the result continual inference. A dedicated deployment gets built around a customer's model, traffic shape and SLO, and the agent keeps re-tuning as load, models or hardware change, on NVIDIA or AMD. Wafer Serverless bills per token from prepaid credits, with OpenAI- and Anthropic-compatible APIs that drop into Claude Code, Cline, OpenHands and similar harnesses. An older flat-rate Wafer Pass from $10 a week still has a page on the site. Artificial Analysis measures Wafer at about 176 tokens per second on GLM-5.3 and 152 on DeepSeek V4.1 Flash, both with up to 1M context. Wafer reports its tuned Qwen 3.5 397B running 2.8x faster than stock SGLang, with GLM 5.1 and DeepSeek V4 Pro each 2x faster than a vLLM baseline.

## Upsides

- Faster open models on the same weights, since most hosts run stock vLLM or SGLang.
- Pay-as-you-go APIs, OpenAI- and Anthropic-compatible, that plug into agentic coding tools.
- Works across NVIDIA and AMD, which helps teams hedge GPU supply.

## Core use cases

- Coding agents that want big open models at interactive speed.
- Dedicated endpoints for teams with a strict latency SLO and no in-house kernel engineers.

## Downsides

- Very young company with a small hosted catalog.
- Speedup claims are self-reported against stock baselines, so compare against tuned hosts like Fireworks before buying.

## At a glance

| Attribute | Value |
|---|---|
| Model access | Open weights |
| Flagship models | GLM-5.3, DeepSeek V4.1 Flash |
| Speed | ~176 tok/s on GLM-5.3; 2–2.8x vs stock vLLM or SGLang (vendor) |
| Price | Per token, prepaid credits |
| Customization | Agent-tuned dedicated deployments |
| Deployment | Serverless, dedicated |
| Long context | Up to 1M (GLM-5.3, DeepSeek V4.1 Flash) |

## FAQ

### What is Wafer?

Wafer builds AI agents that act as GPU performance engineers, then sells inference on the stacks those agents tune. The company came out of Y Combinator's Summer 2025 batch as a "Cursor for CUDA" that turned slow PyTorch into custom kernels. Founders Emilio Andere and Steven Arellano are based in San Francisco. Wafer raised a $4M seed led by Fifty Years in April 2026 and a $40M Series A co-led by Marathon and Chemistry in September 2026 at a $220M post-money valuation, with AMD Ventures and Wing joining. Its agents profile a workload, generate candidate configs across batching, decoding, quantization, engines, kernels and hardware, measure each one and deploy the winner.

### What is Wafer best for?

Coding agents that want big open models at interactive speed; Dedicated endpoints for teams with a strict latency SLO and no in-house kernel engineers.

### How much does Wafer cost?

Wafer pricing at a glance: Per token, prepaid credits. Rates change often, so check Wafer's pricing page before committing.

### How much context does Wafer support?

Wafer's long-context support: Up to 1M (GLM-5.3, DeepSeek V4.1 Flash).

### What are the downsides of Wafer?

Very young company with a small hosted catalog; Speedup claims are self-reported against stock baselines, so compare against tuned hosts like Fireworks before buying.

### What are the best alternatives to Wafer?

Common alternatives include Subconscious, OpenAI, Anthropic, Google Vertex AI, Amazon Bedrock. Each has a head-to-head comparison with Wafer on this site.

## Comparisons

- [Subconscious vs Wafer](https://www.subconscious.dev/compare/subconscious-vs-wafer.md)
- [OpenAI vs Wafer](https://www.subconscious.dev/compare/openai-vs-wafer.md)
- [Anthropic vs Wafer](https://www.subconscious.dev/compare/anthropic-vs-wafer.md)
- [Google Vertex AI vs Wafer](https://www.subconscious.dev/compare/google-vertex-vs-wafer.md)
- [Amazon Bedrock vs Wafer](https://www.subconscious.dev/compare/aws-bedrock-vs-wafer.md)
- [Together AI vs Wafer](https://www.subconscious.dev/compare/together-ai-vs-wafer.md)
- [Fireworks AI vs Wafer](https://www.subconscious.dev/compare/fireworks-vs-wafer.md)
- [Baseten vs Wafer](https://www.subconscious.dev/compare/baseten-vs-wafer.md)
- [Groq vs Wafer](https://www.subconscious.dev/compare/groq-vs-wafer.md)
- [Cerebras vs Wafer](https://www.subconscious.dev/compare/cerebras-vs-wafer.md)
- [DeepInfra vs Wafer](https://www.subconscious.dev/compare/deepinfra-vs-wafer.md)
- [Hugging Face Inference Providers vs Wafer](https://www.subconscious.dev/compare/hugging-face-vs-wafer.md)
- [Modal vs Wafer](https://www.subconscious.dev/compare/modal-vs-wafer.md)
- [Cloudflare Workers AI vs Wafer](https://www.subconscious.dev/compare/cloudflare-workers-ai-vs-wafer.md)
- [xAI vs Wafer](https://www.subconscious.dev/compare/xai-vs-wafer.md)
- [Mistral AI vs Wafer](https://www.subconscious.dev/compare/mistral-ai-vs-wafer.md)
- [DeepSeek vs Wafer](https://www.subconscious.dev/compare/deepseek-vs-wafer.md)
- [Moonshot AI vs Wafer](https://www.subconscious.dev/compare/moonshot-ai-vs-wafer.md)
- [Z.ai vs Wafer](https://www.subconscious.dev/compare/z-ai-vs-wafer.md)
- [Alibaba Cloud vs Wafer](https://www.subconscious.dev/compare/alibaba-cloud-vs-wafer.md)
- [Meta vs Wafer](https://www.subconscious.dev/compare/meta-vs-wafer.md)
- [Cohere vs Wafer](https://www.subconscious.dev/compare/cohere-vs-wafer.md)
- [SambaNova vs Wafer](https://www.subconscious.dev/compare/sambanova-vs-wafer.md)
- [Nebius vs Wafer](https://www.subconscious.dev/compare/nebius-vs-wafer.md)
- [Crusoe vs Wafer](https://www.subconscious.dev/compare/crusoe-vs-wafer.md)
- [fal vs Wafer](https://www.subconscious.dev/compare/fal-vs-wafer.md)
- [Novita AI vs Wafer](https://www.subconscious.dev/compare/novita-ai-vs-wafer.md)
- [Venice vs Wafer](https://www.subconscious.dev/compare/venice-vs-wafer.md)
- [Parasail vs Wafer](https://www.subconscious.dev/compare/parasail-vs-wafer.md)
- [Inference.net vs Wafer](https://www.subconscious.dev/compare/inference-net-vs-wafer.md)
- [GMI Cloud vs Wafer](https://www.subconscious.dev/compare/gmi-cloud-vs-wafer.md)
- [Thinking Machines vs Wafer](https://www.subconscious.dev/compare/thinking-machines-vs-wafer.md)
- [Sail Research vs Wafer](https://www.subconscious.dev/compare/sail-research-vs-wafer.md)
- [Morph vs Wafer](https://www.subconscious.dev/compare/morph-vs-wafer.md)
- [Relace vs Wafer](https://www.subconscious.dev/compare/relace-vs-wafer.md)
- [TypeSafe AI vs Wafer](https://www.subconscious.dev/compare/typesafe-ai-vs-wafer.md)
- [StepFun vs Wafer](https://www.subconscious.dev/compare/stepfun-vs-wafer.md)
- [Runware vs Wafer](https://www.subconscious.dev/compare/runware-vs-wafer.md)
- [StreamLake vs Wafer](https://www.subconscious.dev/compare/streamlake-vs-wafer.md)
- [Wafer vs RunInfra](https://www.subconscious.dev/compare/wafer-vs-runinfra.md)
- [Wafer vs Particle.AI](https://www.subconscious.dev/compare/wafer-vs-particle-ai.md)
- [Wafer vs Luminal](https://www.subconscious.dev/compare/wafer-vs-luminal.md)
- [Wafer vs Infron](https://www.subconscious.dev/compare/wafer-vs-infron.md)
- [Wafer vs Inception](https://www.subconscious.dev/compare/wafer-vs-inception.md)
- [Wafer vs Dolphin](https://www.subconscious.dev/compare/wafer-vs-dolphin.md)

## Sources

- [Wafer](https://www.wafer.ai/)
- [Wafer company info](https://www.wafer.ai/ai-info)
- [Wafer on Y Combinator](https://www.ycombinator.com/companies/wafer)
- [Wafer Serverless docs](https://docs.wafer.ai/serverless)
- [Wafer Series A](https://www.wafer.ai/blog/series-a)
- [Wafer on Artificial Analysis](https://artificialanalysis.ai/providers/wafer)

Pricing and model lineups change often; figures are a snapshot.