Compare startups
Pick two or three companies from the roster to see them side by side. The link in your address bar keeps the comparison, so you can send it to someone.
| Fact | InferenceCallosumRemove | InferenceDoublewordRemove | InferenceFlexAIRemove |
|---|---|---|---|
| Sector | Inference | Inference | Inference |
| What they do | Splits one AI job across whichever models and chips run each part cheapest. | Hosted open-model inference at realtime, async and overnight-batch prices. | One API key for open models that grows from per-token to reserved hardware. |
| Stage | Seed | Not on file | Seed |
| Open roles | 14 | 0 | 0 |
| Tracked since | September 2026 | September 2026 | September 2026 |
| Why we track it | The routing layer across heterogeneous silicon — the software that decides which model runs on which chip, sitting between the chip makers and the teams buying inference. | Open-weight inference sold by latency tier, plus an open-source gateway teams self-host — the layer where a buy-versus-rent decision turns into a per-token bill. | The clearest published attempt to make inference pricing checkable rather than trusted — they write down the rule that sets each rate and ship calculators that show where per-token billing stops being the cheap option. |
Following the roster
Run a startup?
The practice starts by doing the work: what we find goes to you first, and a free weekend on one named problem can follow.