Compute & chips
13 tracked compute & chips, ranked on the AGIX Leaderboard.
| # | Lab | Type | Score | Research roles | SPVs, 12m | Papers, 12m | Models, 12m |
|---|---|---|---|---|---|---|---|
| 4 | CerebrasUS | Compute & chips | 52 | 31 | 9 | – | 17 |
| 21 | Fireworks AIUS | Compute & chips | 26 | 24 | 0 | – | 0 |
| 43 | TenstorrentCanada | Compute & chips | 18 | 5 | 6 | – | – |
| 47 | BasetenUS | Compute & chips | 16 | 13 | 0 | – | – |
| 48 | CoreWeaveUS | Compute & chips | 16 | 11 | 1 | – | – |
| 53 | CrusoeUS | Compute & chips | 14 | 11 | 0 | – | – |
| 54 | GroqUS | Compute & chips | 14 | – | 8 | – | – |
| 57 | EtchedUS | Compute & chips | 12 | 10 | 0 | – | – |
| 58 | LightmatterUS | Compute & chips | 12 | 2 | 5 | – | – |
| 67 | LambdaUS | Compute & chips | 10 | 0 | 5 | – | – |
| 68 | NscaleUK | Compute & chips | 10 | – | 5 | – | – |
| 102 | SambaNovaUS | Compute & chips | 3 | – | 1 | – | 1 |
| 107 | ModularUS | Compute & chips | 2 | – | 0 | – | 3 |
#4CerebrasSunnyvale, US · 2016
Wafer-scale AI chips and fast inference.
31 research roles · 9 SPV filings · 17 models#21Fireworks AIRedwood City, US · 2022Fast inference for open models.
24 research roles#43TenstorrentToronto, Canada · 2016RISC-V AI chips led by Jim Keller.
5 research roles · 6 SPV filings#47BasetenSan Francisco, US · 2019Inference platform for AI models.
13 research roles#48CoreWeaveLivingston, NJ, US · 2017GPU cloud for AI labs (NASDAQ: CRWV).
11 research roles · 1 SPV filings#53CrusoeSan Francisco, US · 2018AI data centres powered by stranded and clean energy.
11 research roles#54GroqMountain View, US · 2016LPU inference chips and cloud.
8 SPV filings#57EtchedSan Jose, US · 2022Sohu, a chip built only for Transformers.
10 research roles#58LightmatterMountain View, US · 2017Photonic interconnects for AI data centres.
2 research roles · 5 SPV filings#67LambdaSan Francisco, US · 2012GPU cloud and servers.
5 SPV filings#68NscaleLondon, UK · 2024European AI cloud and data centres.
5 SPV filings#102SambaNovaPalo Alto, US · 2017Reconfigurable dataflow chips for AI.
1 SPV filings · 1 models#107ModularPalo Alto, US · 2022Mojo language and the MAX inference stack.
3 models