Snapshot
Continue to About ↓- What it does
- General Compute is an AI inference cloud that deploys purpose-built accelerators (Cerebras, SambaNova, Positron AI, d-Matrix) alongside NVIDIA GPUs in US colocation data centers.
More detail
- It sells dedicated bare-metal racks or a tokens-per-second SLA, and also offers an OpenAI-compatible API for open models.
- Best for
- Teams running agents, coding tools or other latency-sensitive workloads that need very high tokens per second per user, and buyers who want dedicated capacity with one contract and one set of SLAs.
More detail
- Also fits companies wanting private or fine-tuned model deployment with model bring-up handled.
- Pricing
- Dedicated capacity and private model deployment are priced per contract on a rate card; no public price list.
More detail
- The rate depends on workload shape, chip, model and term, and longer terms get lower rates.
- Exact figures, minimum terms and ancillary costs (egress, setup, deposits) are undisclosed and onl
- Main limitation
- Buyers who need inference capacity live before Q1 2027, since its deployments go live then.
More detail
- Also a poor fit for teams needing non-US colocation, guaranteed SLAs on pay-as-you-go, or a public price list before committing.
- Record
- Living and monitored
- Reviewed
- Evidence
- Public web research
- Machine access
- Structured profile ↗
About
- General Compute is an AI inference cloud that deploys purpose-built accelerators (Cerebras, SambaNova, Positron AI, d-Matrix) alongside NVIDIA GPUs in US colocation data centers.
- It sells dedicated bare-metal racks or a tokens-per-second SLA, and also offers an OpenAI-compatible API for open models.
What it helps with
Capabilities
A practical view of what is available and where limitations apply.
| Offering | Available | Detail |
|---|---|---|
| Disaggregated inference (prefill on GPU, decode on ASIC) | Yes | Prefill on NVIDIA GPUs, decode on Cerebras, SambaNova, Positron AI or d-Matrix |
| Cerebras CS-3 tier | Yes | 2 x CS-3 per 48U rack, 47 kW average / 55 kW rated, water-cooled, 44 GB on-wafer SRAM, 1k+ tok/s/user on multi-trillion-parameter models |
| Bare metal or OpenAI-compatible endpoint | Yes | Root on dedicated machines, or an OpenAI-compatible endpoint |
| OpenAI-compatible API | Yes | REST API plus Node.js and Python SDKs |
| API features (tool calling, JSON mode, streaming, reasoning) | Yes | Tool/function calling, JSON mode, reasoning-ready models with extended thinking timeouts, streaming |
| Custom checkpoints (LoRA/GGUF/full fine-tune) | Yes | Bring your own weights, exposed via a private model ID |
| Public price list for dedicated capacity | No | Priced per contract on a rate card; figures undisclosed until after a call |
| Contractual SLA on pay-as-you-go | Limited | PAYG plans inherit best-effort reliability; contractual SLAs are Enterprise-only |
| Live capacity before Q1 2027 | No | Deployments go live from Q1 2027 |
| Non-US colocation | Limited | US colocation only for dedicated capacity; other API regions Enterprise-only |
Bot trust
How confidently AI systems can verify this profile from public evidence.
—of 100
ChatGPT—/100
Claude—/100
This profile has not completed a Bot Score assessment yet.
Evidence
Where this record came from, when it was reviewed, and how confidently it can be used.
- Verification
- Scribble researched these facts from the linked public sources.
- The owner has not manually confirmed every field.
- Last reviewed
- Primary source
- generalcompute.com
Technical profile details
Record ID
0587bf4e-6a0f-4fba-b366-ebf869ac8dd6Canonical URLhttps://profiles.scribble.network/listing/general-compute-0587bf4e