vs
Nebius vs Wafer
A scaled European cloud measured among the top hosts on throughput against a young company whose agents tune inference stacks per workload.
By The Subconscious Team · Updated
Nebius vs Wafer: key differences
Wafer's pitch is that most hosts run stock vLLM or SGLang, and its agents can do better. They profile a workload, test configs across batching, quantization, engines, kernels and hardware, and deploy the winner, then keep re-tuning as traffic changes on NVIDIA or AMD. Wafer reports its tuned Qwen 3.5 397B at 2.8x stock SGLang and GLM 5.1 and DeepSeek V4 Pro each 2x a vLLM baseline. Those are self-reported against untuned baselines. Nebius is a larger, established cloud that Artificial Analysis has independently measured among the top hosts on throughput, with speculative decoding on dedicated endpoints.
The two also package access differently. Wafer Serverless bills per token from prepaid credits, aimed at Claude Code, Cline and OpenHands users. Nebius bills per token from $0.06 per million input and adds GPU rental and EU or US placement. Wafer is very young with a small hosted catalog, so Nebius is the safer bet for breadth, residency and scale. Wafer is worth testing for a strict latency SLO on a big open model when no one in-house tunes kernels.
What Nebius and Wafer do
Nebius
Nebius is an Amsterdam-headquartered AI cloud and the strongest European alternative to the US hyperscalers. It sells raw NVIDIA GPU compute, from H100s at $4.50 an hour on demand or from $0.79 on dynamic spot pricing, up to GB300 NVL72 racks, and it has begun adding Vera Rubin. On-demand GPU prices rose on October 1, 2026, and preemptible VMs moved to demand-set spot prices on October 8. Hyperscale buyers back it: a Microsoft capacity deal worth about $17.4B in September 2025, then a Meta agreement worth up to about $27B in March 2026.
Example models: DeepSeek V4 Pro, Kimi K3, MiniMax M3
Full Nebius profileWafer
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.
Example models: GLM-5.3, DeepSeek V4.1 Flash
Full Wafer profileShould you choose Nebius or Wafer?
Nebius
Choose Nebius for
- Independently measured throughput across a 60+ model catalog
- EU residency and a 99.9% dedicated-endpoint SLA
- Moving from tokens into GPU training on one account
Wafer
Choose Wafer for
- Per-token access for agentic coding harnesses
- Dedicated endpoints re-tuned continuously for a strict latency SLO
- Teams hedging GPU supply across NVIDIA and AMD
Nebius vs Wafer at a glance
| Attribute | Nebius | Wafer |
|---|---|---|
| Model access | Open weights, 60+ models | Open weights |
| Flagship models | DeepSeek V4 Pro, Kimi K3, MiniMax M3, GLM | GLM-5.3, DeepSeek V4.1 Flash |
| Speed | Among top hosts on throughput | ~168 tok/s on GLM-5.3; 2–2.8x vs stock vLLM or SGLang (vendor) |
| Price | From $0.06 per 1M input | Per token, prepaid credits |
| Customization | Serve uploaded fine-tunes | Agent-tuned dedicated deployments |
| Deployment | Token Factory, dedicated, raw GPUs | Serverless, dedicated |
| Long context | Varies by model | Up to 1M (GLM-5.3, DeepSeek V4.1 Flash) |
Frequently asked questions
What is the difference between Nebius and Wafer?
A scaled European cloud measured among the top hosts on throughput against a young company whose agents tune inference stacks per workload.
When should I choose Nebius over Wafer?
Independently measured throughput across a 60+ model catalog; EU residency and a 99.9% dedicated-endpoint SLA; Moving from tokens into GPU training on one account.
When should I choose Wafer over Nebius?
Per-token access for agentic coding harnesses; Dedicated endpoints re-tuned continuously for a strict latency SLO; Teams hedging GPU supply across NVIDIA and AMD.
Is Nebius or Wafer cheaper?
Nebius: From $0.06 per 1M input. Wafer: Per token, prepaid credits. The cheaper choice depends on the model and workload.
Which has more context, Nebius or Wafer?
Nebius: Varies by model. Wafer: Up to 1M (GLM-5.3, DeepSeek V4.1 Flash).
Related comparisons
Running long-horizon agents?
If your agents run past 200K tokens, compare both against Subconscious. Our inference stack treats a long-horizon trace as the primary workload, so speed, cost, and accuracy hold up deep into the trace.