Register for a premium account to gain access to Sterling AI.
Get StartedThings you can ask Sterling:
| Anchor Model / Customer | Role on Replicate | Published Price (where applicable) | Significance |
|---|---|---|---|
| Stable Diffusion (Stability AI) | Historic anchor image model | Per-second GPU billing | Made Replicate the canonical hosted text-to-image endpoint in 2022; one of the largest sources of early run volume |
| FLUX (Black Forest Labs) | Anchor image model 2024-2025 | Flux 1.1 Pro $0.04 / image; Flux Dev $0.025 / image | Replicate became the default place to run FLUX; major source of platform run volume into the acquisition |
| Ideogram V3 | High-end image model | $0.09 / output image | Premium per-image tier; illustrates per-output monetization above raw GPU cost |
| Wan 2.1 (video) | Anchor video model | $0.09 / second of 480p output | Video is the fastest-growing and most compute-intensive run category |
| Llama family (Meta) | Anchor open-source LLM | Per-second GPU billing (often via Cog + vLLM) | Drove the 2023-2024 open-source LLM wave on the platform |
| Claude 3.7 Sonnet (Anthropic) | Hosted proprietary LLM | $3.00 / million input tokens; $0.015 / thousand output tokens | Token-priced model offered alongside open-source catalog |
| Whisper (OpenAI, open-source) | Anchor audio model | Per-second GPU billing | Canonical hosted speech-to-text endpoint for the long tail |
| BuzzFeed, Character.ai (early) | Early anchor consumer customers | Undisclosed | First high-traffic consumer apps standardized on Replicate |
| Unsplash | Named customer (per Sacra) | Undisclosed | Platform processes 'tens of millions of calls' for image and media use cases |
2021-2026-H1
| Round (date, lead) | Amount Raised | Post-Money Valuation | Notable Investors |
|---|---|---|---|
| Seed (2020-09) | $5.3M | n/a | Y Combinator, Andreessen Horowitz |
| Series A (2023-02, Andreessen Horowitz) | $12.5M | n/a (undisclosed) | Andreessen Horowitz (lead), Y Combinator, Sequoia Capital, angels |
| Series B (2023-12, Andreessen Horowitz) | $40M | $350M | Andreessen Horowitz (lead), Sequoia Capital, NVentures (Nvidia), Y Combinator, Heavybit |
| Acquisition by Cloudflare (2025-11) | Undisclosed (RUMORED $350M to $550M, unconfirmed) | n/a | Cloudflare (NYSE: NET); folds Replicate into Workers AI |
| Total raised (pre-acquisition) | ~$58M | n/a | Capital-light for an AI-infra company; a16z anchored all three priced rounds |
| Hardware Tier | Price / Second (USD) | Price / Hour (USD) | Notes |
|---|---|---|---|
| CPU (small) | $0.000025 | $0.09 | Lightweight pre/post-processing, no GPU |
| CPU | $0.000100 | $0.36 | Cheapest compute option on the platform |
| Nvidia T4 (16GB) | $0.000225 | $0.81 | Entry GPU for small image and audio models |
| Nvidia L40S (48GB) | $0.000975 | $3.51 | Mid-tier image and video inference |
| Nvidia A100 (80GB) | $0.001400 | $5.04 | Workhorse for large diffusion and LLM models |
| Nvidia H100 (80GB) | $0.001525 | $5.49 | Highest-throughput single-GPU tier |
| 2x Nvidia L40S | $0.001950 | $7.02 | Multi-GPU image and video workloads |
| 2x Nvidia A100 (80GB) | $0.002800 | $10.08 | Larger-memory LLM and video models |
| 2x Nvidia H100 | $0.003050 | $10.98 | Requires committed (reserved) contract |
| 4x Nvidia A100 (80GB) | $0.005600 | $20.16 | Requires committed (reserved) contract |
2022-Nov 2025

$PRIVATE
Largest AI lab by revenue. Approximately $13B annualized run-rate revenue (mid-2025) from a mix of ChatGPT consumer subscriptions, ChatGPT Enterprise / Team, the OpenAI API, and Microsoft Azure OpenAI revenue share. ChatGPT reaches roughly 800M weekly active users (late 2025). Builder of the GPT family (GPT-4o, GPT-4.5, GPT-5, o-series reasoning models) and Sora video. Microsoft is the dominant compute and distribution partner; SoftBank led the $40B March 2025 primary round at $300B.