# Fireworks AI vs Replicate: who uses which

On usedby, 59 companies use Fireworks AI and 14 use Replicate. 2 companies use both, including Figma and Timbal. Counts include every proof level, job-post signals included; each company keeps its proof level on each side. Latest check 2026-11-22: 73 of 73 public lines checked.

## Side by side

|  | [Fireworks AI](https://www.usedby.ai/tools/fireworks-ai.md) | [Replicate](https://www.usedby.ai/tools/replicate.md) |
|---|---|---|
| Public customers | 59 | 14 |
| Confirmed | 14 | 0 |
| Observed | 23 | 7 |
| Signals (job posts) | 0 | 0 |
| Top industries of customers (every proof level) | DevTools & Infrastructure (19), Artificial Intelligence (17), IT Services & System Integration (4) | Artificial Intelligence (6), Marketing Tech (2), DevTools & Infrastructure (1) |
| Category | AI inference & GPU cloud | AI inference & GPU cloud |
| Security | SOC 2 Type II, HIPAA, ISO 27001, ISO 27701, ISO 42001 ([trust.fireworks.ai](https://trust.fireworks.ai/), checked 2026-11-13) | — |
| Data processing terms | A data processing agreement is available. ([trust.fireworks.ai](https://trust.fireworks.ai/), checked 2026-11-13) | — |
| Pricing | Serverless · From $0.05 / million input tokens ([fireworks.ai](https://fireworks.ai/pricing), checked 2026-11-13) | CPU: $0.000100 sec; Nvidia T4 GPU: $0.000225 sec; Nvidia A100 (80GB) GPU: $0.001400 sec ([replicate.com](https://replicate.com/pricing), checked 2026-11-16) |

## Companies that use both

Every proof level counts; the proof level of each side is shown, strongest proofs first. The first public proof is not an adoption date.

| Company | Proof, Fireworks AI | Proof, Replicate | First public proof, Fireworks AI | First public proof, Replicate |
|---|---|---|---|---|
| [Figma](https://www.usedby.ai/companies/figma.md) | Declared by the company | Declared by the company | On or before November 2026 | On or before November 2026 |
| [Timbal](https://www.usedby.ai/companies/timbal.md) | Declared by the company | Declared by the company | On or before November 2026 | On or before November 2026 |

## Published results

Reviewed stories only; usedby did not measure the figures.

- Fireworks AI: [Cursor](https://www.usedby.ai/case-studies/cursor-fireworks-ai-2.md): **~13x**: “enabling them to ~13x speedup over vanilla inference using Llama-3-70b” (published 2024-06-23)
- Fireworks AI: [Cresta](https://www.usedby.ai/case-studies/cresta-fireworks-ai.md): **100x**: “Cresta achieves up to a 100x cost reduction compared to GPT-4” (published 2024-12-08)

---
Source: https://www.usedby.ai/compare/fireworks-ai-vs-replicate · How we check: https://www.usedby.ai/methodology
