General Instinct GTM brain

General Instinct builds post-training and deployment infrastructure for teams running physical AI models on robots and edge devices.

category

Physical AI deployment infrastructure

market profile

  • text: Robot developers must move models from lab demonstrations to reliable operation under constrained edge hardware, latency and connectivity.
  • text: General Instinct offers a model runtime and post-training infrastructure rather than a specific robot platform.

personas

  • title: Head of robotics or robot learning
    pains:
    • Promising robot policies fail to transfer reliably to real environments.
    • Post-training and deploying across machines becomes harder as fleets grow.
  • title: Edge AI inference engineer
    pains:
    • Large models exceed device memory or latency budgets.
    • Hardware-specific optimizations and offline operation require specialized engineering.
  • title: Robotics product or deployment lead
    pains:
    • Cloud-dependent inference can be expensive or unavailable at the deployment site.
    • A model that runs in a demo must be validated on target hardware before rollout.

icp

  • text: Robot, drone and physical AI developers deploying vision or action models on resource-constrained edge devices; size is unspecified.
  • text: Teams moving from prototype to real deployments or scaling robot policies across a fleet are plausible early buyers.

message pillars

  • text: Provide the infrastructure needed to deploy physical AI.
  • text: Make robot post-training reliable at scale.
  • text: Focus on reliable deployment for large robot programs.

objections

  • objection: Can the runtime meet our latency and accuracy requirements?
    response: Benchmark the actual model on target hardware with representative sensors and an agreed latency budget.
  • objection: Will it fit our robot and model stack?
    response: Review supported models, device architectures and integration with existing controls in a technical pilot.

competitors

  • name: ZETIC
    website: https://zetic.ai/
    why: Offers hardware-aware inference optimization for robots.
  • name: Viam
    website: https://www.viam.com/
    why: Offers physical AI development and deployment infrastructure, with broader robot management.
  • name: NVIDIA
    why: TensorRT and Jetson provide an alternative edge inference stack and may also complement General Instinct.

influencers

  • name: Bill Jiao
    url: https://luma.com/ghtt4qce
    why: Founder speaks publicly about deploying frontier models on robots and edge devices.

keywords

  • robot model deployment
  • physical AI infrastructure
  • robot post training
  • edge vision model inference
  • robot AI model optimization
  • robot fleet deployment

voice

  • Technical
  • Concise
  • Reliability-focused

signals

  • text: A robotics team is trying to fit large vision or action models on constrained edge hardware.
  • text: A team is scaling from robot pilots to repeated real-world deployment.

stack

  • InstinctFlash
  • NVIDIA Jetson
  • CUDA
  • Metal
  • ARM NEON

pricing

model: unknown

motion

hybrid

channels

  • website
  • open source
  • X
  • LinkedIn
  • robotics events

confidence

medium

Founders

  • name: Bill Jiao
    title: Founder/CEO
    role: CEO
    bio: worked on Siemens' first multimodal foundation model, secured 2 patents on tardigrade proteins as a junior in high school, now building edge intelligence for physical AI.
  • name: Guanming Wang
    title: Founder/CTO
    role: CTO
    bio: Ex Google DeepMind reinforcement learning researcher with robotics research published at ICRA at age 17. Previously conducted robotics research at UCL robotics institute and Tsinghua University.

Back to YC companies