Top accelerator startups

Startups14,448Programs29

14,448 startups · 29 programs

Harbor AI

The operating layer that keeps GPU fleets productive.

Websitespeedrun page

Why it ranks

How the rank works
3signal of 100 · rising 2
  • went through speedrunProgram · this site+3

A sum of public facts, not a judgement and not investment advice.

Competitors

AIOpereitShared: operating layerSeedcamp 2026Startup →AIThunder ComputeShared: gpu fleet, computeYC S24San FranciscoStartup →AIPlena HealthShared: operating layer, keep, runningYC X26San FranciscoStartup →AISekiloShared: operating layer, starting, buildingYC W22JakartaStartup →OtherLambdaShared: gpu, inference, computeStartup →
Companies closest to Harbor AI
Company
OpereitShared: operating layerSeedcamp 2026Not statedAI
Thunder ComputeShared: gpu fleet, computeYC S24ActiveAI, Data, Devtools
Plena HealthShared: operating layer, keep, runningYC X26ActiveAI
SekiloShared: operating layer, starting, buildingYC W22ActiveAI
Lambdanot in a program hereShared: gpu, inference, computeElsewhereOther

Adjacent

AIProphecta CorpShared: fleetAlchemist Class 36Walnut CreekStartup →AInCompass TechnologiesShared: gpu, longer, inferenceYC W24New York CityStartup →AIZibra AIShared: ai infra, gpu, layerspeedrun SR001LewesStartup →OtherPauaShared: fleet, startingSeedcamp 2021Startup →AIEvokonnect Pte Ltd (Just Ping)Shared: fleetAntler 2023SingaporeStartup →
Companies next to Harbor AI
Company
Prophecta CorpShared: fleetAlchemist Class 36Not statedAI
nCompass TechnologiesShared: gpu, longer, inferenceYC W24ActiveAI, Devtools
Zibra AIShared: ai infra, gpu, layerspeedrun SR001Not statedAI
PauaShared: fleet, startingSeedcamp 2021Not statedOther
Evokonnect Pte Ltd (Just Ping)Shared: fleetAntler 2023Not statedAI

Also from speedrun SR007 in AI

All of speedrun SR007
AIAgentCatSession replay and observability for AI agents across MCP, Claude Connectors, and ChatGPT Apps.speedrun SR007New York CityStartup →AIChatOverflowThe network where AI agents learn from each otherspeedrun SR007San FranciscoStartup →AICodosThe virtual Chief AI Officer that turns AI into P&L impact. 7-digit live ARR, 50% MoM, ex-McKinsey and Meta ML team.speedrun SR007San FranciscoStartup →AICube ComputerPowerful, one-click cloud dev boxes for running agents.speedrun SR007New York CityStartup →AICybernetic PhysicsAutonomous robots to operate data centers.speedrun SR007StanfordStartup →AIDelightful AIAI agent worlds for science and healthcare.speedrun SR007San FranciscoStartup →
Company
AgentCatSession replay and observability for AI agents across MCP, Claude Connectors, and ChatGPT Apps.Not statedAI, Devtools
ChatOverflowThe network where AI agents learn from each otherNot statedAI, Devtools
CodosThe virtual Chief AI Officer that turns AI into P&L impact. 7-digit live ARR, 50% MoM, ex-McKinsey and Meta ML team.Not statedAI
Cube ComputerPowerful, one-click cloud dev boxes for running agents.Not statedAI, Devtools
Cybernetic PhysicsAutonomous robots to operate data centers.Not statedAI
Delightful AIAI agent worlds for science and healthcare.Not statedAI, Data

Programs

Programs Harbor AI went through
ProgramCohortSource
a16z speedrunSR007speedrun Page

About Harbor AI

We’re building Harbor to be the operating layer that keeps GPU fleets productive. A fleet can be completely online and still be wasting compute. Training takes longer, or an inference service handles fewer requests than it should. We’re starting with software for AI teams running dedicated GPUs that figures out why workloads are failing or slowing down, makes supported fixes with approval, and checks whether performance recovers. We’re building a context engine that keeps track of the setup, what changed and what worked before, and developing models that use that understanding to predict which changes will help. Long term, a team should be able to tell Harbor what work needs to get done and what performance it needs. Harbor would figure out how to run that work across compute they own or rent, balancing speed, reliability and cost. As workloads and requirements change, Harbor would keep learning and adjusting how the fleet runs.

Category
AI
Location
San Francisco, California, United States of America
Checked
10h ago

Tags: AI Infra

Harbor AI: questions

Which accelerators did Harbor AI go through?

Harbor AI is listed by a16z speedrun (SR007).

Where is Harbor AI based?

Harbor AI lists San Francisco, California, United States of America as its location on a16z speedrun's page.