OneTriangle

Inference company that moves the KV cache from a small prefill model to a large decode model so the large model skips most prefill compute, and serves managed open-weight models such as DeepSeek, Llama and Qwen at per-token prices.

StageUndisclosed

Founded2026

HeadquartersSan Francisco, USA

StatusPrivate

Founded — ycombinator.com · 20 Oct 2026

onetriangle.ai

Product

Selected offerings — the AI products recorded here, not the vendor's full catalogue.

OneTriangle Inference Model API Sold on its own

Inference service that prefills long inputs on a small model and decodes on a large one by transferring the KV cache between them, falling back to ordinary prefill when a model pair does not meet quality and latency gates, and also serves open-weight models such as DeepSeek V4 Flash, Llama 3.3 70B and Qwen3 variants on managed GPUs; priced per million input and output tokens, with a demo on request.

Similar products 81

No sourced competitor yet: no filing, vendor document or measured-market report read so far names a rival for this company, so none is asserted. The companies below are the closest by what their products do.

How this list is ranked ↗

All 81Capability overlap 76Shared product type 24Same company status 61

IsoquantUndisclosedOneTriangle Inference and Isoquant both list model hosting; both also serve buyers who need to serve a model in production and track and cut what we spend on AI models; scale not recorded for either.2 of 2 capabilitiesShared product typeCompare O

OllamaEarly-stageOneTriangle Inference and Ollama Cloud both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either.2 of 2 capabilitiesShared product typeCompare R

RunAnywhereUndisclosedOneTriangle Inference and Wally both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either.2 of 2 capabilitiesShared product typeCompare B

BasetenLate-stageOneTriangle Inference and Baseten Model APIs both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare NG

Nebius GroupPublicOneTriangle Inference and Nebius Token Factory both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare R

ReplicateUndisclosedOneTriangle Inference and Replicate both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either; ships infrastructure service rather than the same layer.2 of 2 capabilitiesDifferent layerCompare FA

Fireworks AILate-stageOneTriangle Inference and Fireworks Inference both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare TA

Together AILate-stageOneTriangle Inference and Together Dedicated Model Inference both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare R

RunPodLate-stageOneTriangle Inference and Runpod Hub both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare L

LambdaLate-stageOneTriangle Inference and Lambda Inference both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either.2 of 2 capabilitiesShared product typeCompare V

VMwareAcquiredOneTriangle Inference and VMware Cloud Foundation Private AI Services both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; VMware is acquired, with no independent scale; ships infrastructure service rather than the same layer.2 of 2 capabilitiesDifferent layerCompare O

OpenRelayUndisclosedOneTriangle Inference and OpenRelay both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either; ships infrastructure service rather than the same layer.2 of 2 capabilitiesDifferent layerCompare HC

Huawei CloudUndisclosedOneTriangle Inference and MaaS both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either.2 of 2 capabilitiesShared product typeCompare M

ModalLate-stageOneTriangle Inference and Modal Inference both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships developer tool and infrastructure service rather than the same layer.2 of 2 capabilitiesDifferent layerCompare C

ConiferUndisclosedOneTriangle Inference and Conifer both list model hosting; both also serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service, application and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare M

ModularAcquiredOneTriangle Inference and Modular Cloud both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; Modular is acquired, with no independent scale; ships developer tool and infrastructure service rather than the same layer.2 of 2 capabilitiesDifferent layerCompare KE

Kakao EnterpriseUndisclosedOneTriangle Inference and Kubeflow both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either; ships developer tool and platform rather than the same layer.2 of 2 capabilitiesDifferent layerCompare C

CrusoeLate-stageOneTriangle Inference and Crusoe Managed Inference both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships API service, developer tool and 2 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare D

DatabricksLate-stageOneTriangle Inference and Databricks Model Serving both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare HF

Hugging FaceLate-stageOneTriangle Inference and Spaces both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships AI agent, API service and 3 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare C

CivitaiEarly-stageOneTriangle Inference and Civitai API both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either; ships API service and platform rather than the same layer.2 of 2 capabilitiesDifferent layerCompare E

ElasticPublicOneTriangle Inference and Elastic Inference Service both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare W&

Weights & BiasesAcquiredOneTriangle Inference and Serverless Inference both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; Weights & Biases is acquired, with no independent scale; ships API service, developer tool and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare LS

LM StudioUndisclosedOneTriangle Inference and LM Studio both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either; ships AI agent, application and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare F

FalUndisclosedOneTriangle Inference and fal both serve buyers who need to serve a model in production; scale not recorded for either.2 of 2 capabilitiesShared product typeCompare D

DecartLate-stageOneTriangle Inference and Cogito both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare MP

Meta PlatformsPublicOneTriangle Inference and Meta Model API both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare H

HeliconeAcquiredOneTriangle Inference and Helicone AI Gateway both list model hosting; OneTriangle's scale not recorded; Helicone is acquired, with no independent scale; ships API service and platform rather than the same layer.1 of 2 capabilitiesDifferent layerCompare NC

NAVER CloudUndisclosedOneTriangle Inference and ML expert Platform both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either; ships API service, application and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare C

CohereLate-stageOneTriangle Inference and Model Vault both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare VE

Volcano EngineUndisclosedOneTriangle Inference and Machine Learning Platform (机器学习平台) both serve buyers who need to serve a model in production; scale not recorded for either.1 of 2 capabilitiesShared product typeCompare N

NVIDIAPublicOneTriangle Inference and NVIDIA NIM both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare EL

Experiential LabsUndisclosedOneTriangle Inference and Experiential Labs Gateway both serve buyers who need to serve a model in production and track and cut what we spend on AI models; scale not recorded for either; ships API service and developer tool rather than the same layer.1 of 2 capabilitiesDifferent layerCompare T

TencentPublicOneTriangle Inference and TokenHub both list model hosting; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare A

AIHubMixUndisclosedOneTriangle Inference and LLM Router both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare I

InfronUndisclosedOneTriangle Inference and Infron both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare TT

The Token CompanyUndisclosedOneTriangle Inference and The Token Company LLM input compression both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare V

VercelLate-stageOneTriangle Inference and Vercel AI Gateway both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships AI agent, API service and 2 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare H

HopperUndisclosedOneTriangle Inference and Hopper both serve buyers who need to serve a model in production; scale not recorded for either.1 of 2 capabilitiesShared product typeCompare A

AMDPublicOneTriangle Inference and AMD Enterprise AI Reference Stack both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships application, hardware and 2 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare DT

Dell TechnologiesPublicOneTriangle Inference and Dell Enterprise Hub both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships application, data service and 3 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare LG

LLM GatewayUndisclosedOneTriangle Inference and LLM Gateway both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare T

TrimioUndisclosedOneTriangle Inference and Trimio both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare Z

ZenMuxUndisclosedOneTriangle Inference and ZenMux both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare H

HuaweiUndisclosedOneTriangle Inference and MindIE both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; Huawei's scale not comparable: only a revenue estimate, no market cap or valuation; ships application, developer tool and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare L

LiteLLMEarly-stageOneTriangle Inference and Auto Router both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships AI agent, application and 1 more rather than the same layer.1 of 2 capabilitiesDifferent layerCompare ND

Not DiamondUndisclosedOneTriangle Inference and Not Diamond Code both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service and developer tool rather than the same layer.1 of 2 capabilitiesDifferent layerCompare T

TrueFoundryEarly-stageOneTriangle Inference and AI Deployment Platform both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either; ships API service, agent platform and 2 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare L

LangDBUndisclosedOneTriangle Inference and LangDB AI Gateway both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare R

RespanEarly-stageOneTriangle Inference and Respan AI Gateway both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service and infrastructure service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare K

KuaishouPublicOneTriangle Inference and Vanchin both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare T

TouchmarkUndisclosedOneTriangle Inference and Touchmark both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships platform rather than the same layer.1 of 2 capabilitiesDifferent layerCompare B

BelvedirUndisclosedOneTriangle Inference and Belvedir both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships platform rather than the same layer.1 of 2 capabilitiesDifferent layerCompare O

OrchestraUndisclosedOneTriangle Inference and Orchestra both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships platform rather than the same layer.1 of 2 capabilitiesDifferent layerCompare O

Orq.aiUndisclosedOneTriangle Inference and AI Gateway both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare CA

Concentrate AIUndisclosedOneTriangle Inference and LLM Gateway both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare Q

QualcommPublicOneTriangle Inference and Qualcomm AI Inference Suite both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships application, developer tool and 2 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare FA

Future AGIUndisclosedOneTriangle Inference and Command Center both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare J

JFrogPublicOneTriangle Inference and JFrog AI Catalog both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either; ships developer tool, infrastructure service and 1 more rather than the same layer.1 of 2 capabilitiesDifferent layerCompare A

AlibabaPublicOneTriangle Inference and ModelScope both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare O

OpenRouterLate-stageOneTriangle Inference and OpenRouter both serve buyers who need to serve a model in production; OneTriangle's scale not recorded.1 of 2 capabilitiesShared product typeCompare L

LangWatchUndisclosedOneTriangle Inference and LangWatch AI Gateway both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships AI agent, API service and 2 more rather than the same layer.1 of 2 capabilitiesDifferent layerCompare N

nRouterUndisclosedOneTriangle Inference and nRouter LLM Gateway both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships API service rather than the same layer.Different layerCompare A

AmazonPublicOneTriangle Inference and Amazon SageMaker AI both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded.2 of 2 capabilitiesShared product typeCompare G

GitLabPublicOneTriangle Inference and GitLab Duo Agent Platform Self-Hosted both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships AI agent, agent platform and 4 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare CC

Cactus ComputeUndisclosedOneTriangle Inference and Cactus both serve buyers who need to serve a model in production; scale not recorded for either; ships infrastructure service and platform rather than the same layer.1 of 2 capabilitiesDifferent layerCompare Y

YandexPublicOneTriangle Inference and Yandex DataSphere both list model hosting; both also serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships API service, application and 2 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare L

LatitudeUndisclosedOneTriangle Inference and Latitude both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships platform rather than the same layer.Different layerCompare S

SpendlineUndisclosedOneTriangle Inference and Spendline both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships platform rather than the same layer.Different layerCompare C

CoreWeavePublicOneTriangle Inference and CoreWeave Dedicated Inference both serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships AI agent, API service and 4 more rather than the same layer.1 of 2 capabilitiesDifferent layerCompare V

Vast.aiUndisclosedOneTriangle Inference and Vast.ai Serverless both serve buyers who need to serve a model in production; scale not recorded for either; ships infrastructure service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare M

MossUndisclosedOneTriangle Inference and Moss Inference both serve buyers who need to serve a model in production; scale not recorded for either; ships AI agent and infrastructure service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare BA

Blackbox AIUndisclosedOneTriangle Inference and Blackbox Enterprise Inference both list model hosting; both also serve buyers who need to serve a model in production; scale not recorded for either; ships AI agent, API service and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare AS

AI SpendOpsUndisclosedOneTriangle Inference and AI SpendOps both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships infrastructure service rather than the same layer.Different layerCompare M

MicrosoftPublicOneTriangle Inference and Foundry Models both list model hosting; OneTriangle's scale not recorded; ships AI agent, API service and 6 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare U

UnslothUndisclosedOneTriangle Inference and Unsloth both serve buyers who need to serve a model in production; scale not recorded for either; ships developer tool rather than the same layer.1 of 2 capabilitiesDifferent layerCompare NC

Nodus ComputeUndisclosedOneTriangle Inference and Nodus both serve buyers who need to serve a model in production; scale not recorded for either; ships infrastructure service rather than the same layer.1 of 2 capabilitiesDifferent layerCompare CM

Comet MLEarly-stageOneTriangle Inference and Cost Intelligence both serve buyers who need to track and cut what we spend on AI models; scale not recorded for either; ships AI agent, application and 2 more rather than the same layer.Different layerCompare NR

Nous ResearchEarly-stageOneTriangle Inference and Hermes Cloud both list model hosting; scale not recorded for either; ships AI agent, API service and 2 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompare L

LovableLate-stageOneTriangle Inference and Lovable Cloud both list model hosting; OneTriangle's scale not recorded; ships platform rather than the same layer.1 of 2 capabilitiesDifferent layerCompare U

UniphoreLate-stageOneTriangle Inference and Uniphore Model Layer both serve buyers who need to serve a model in production; OneTriangle's scale not recorded; ships application, data service and 1 more rather than the same layer.1 of 2 capabilitiesDifferent layerCompare

Show all 81 related companiesShow fewer

Closest alternatives are pairs a vendor page itself compares. Similar products are matched from what each product does; nothing in that group is asserted as a rivalry.