General Intuition Raises $134M Seed Round for Spatial AI Agents. Yes, You Read That Right - $134M Seed.

A gaming company spinoff just raised $133.7 million in seed funding.

Let that sink in. Seed round. Not Series A. Not Series B. A seed round that's larger than most companies' entire fundraising history.

General Intuition, fresh out of stealth and spun off from gaming platform Medal, pulled off one of the biggest seed rounds in AI history. The company is building spatial-temporal AI agents - basically AI that understands how objects move and interact in physical 3D space.

They're training these agents on massive video game datasets. The target applications? Rescue drones, industrial robotics, and autonomous systems that need to navigate complex real-world environments.

Translation: VCs just bet $134 million that video games will teach AI to replace human operators in physical industries.

What Happened: Gaming AI Goes Physical

On November 3, 2025, General Intuition announced its launch with a $133.7 million seed round led by Khosla Ventures, with participation from General Catalyst and Raine.

The company spun out of Medal, a gaming platform that's been collecting massive amounts of gameplay data. Now they're using that data to train AI agents that understand spatial reasoning, physics, and temporal dynamics - how things move and interact over time in 3D space.

Why video games? Because games are perfect training grounds for AI that needs to understand physical space:

• Simulated physics: Games model gravity, momentum, collision, friction
• Complex environments: Buildings, terrain, obstacles, dynamic objects
• Real-time decision-making: React to changing conditions instantly
• Goal-oriented behavior: Navigate to objectives, avoid hazards, optimize paths
• Massive scale data: Millions of hours of gameplay showing how to navigate 3D spaces

Traditional AI models like GPT are trained on text and images. They're 2D. They don't understand depth, movement, physics, or how objects interact in space over time. General Intuition is building AI that does.

Why This Matters: Next-Gen AI Architecture

Here's why VCs dropped $134 million on a seed round:

Current AI is really good at digital tasks - writing, coding, analyzing data, generating images. That's why it's eating jobs in content creation, customer service, and knowledge work.

But AI has struggled with physical tasks that require spatial reasoning. Warehouse robots need carefully mapped environments. Drones crash into obstacles. Industrial robots need extensive programming for each specific task.

The limitation isn't hardware. It's software. Current AI doesn't really understand how the physical world works.

General Intuition's approach - training AI on millions of hours of gameplay where agents navigate complex 3D environments, avoid obstacles, plan paths, and accomplish objectives - could be the breakthrough that makes AI actually useful in physical industries.

What Gets Automated

If this tech works as intended, here's what becomes automatable:

Rescue and Emergency Response:

• Search and rescue drone operations (currently requires skilled human pilots)
• Disaster site navigation and assessment
• Emergency supply delivery to hard-to-reach areas
• Hazardous environment inspection

Industrial Operations:

• Warehouse navigation without fixed tracks or rails
• Construction site material handling
• Facility inspection and maintenance
• Complex assembly tasks currently requiring human dexterity

Transportation and Logistics:

• Last-mile delivery (navigating sidewalks, stairs, obstacles)
• Port and rail yard operations
• Agricultural equipment operation
• Mining and resource extraction

All these industries currently employ human operators because existing automation couldn't handle the complexity of dynamic, unstructured environments. Spatial AI changes that equation.

The Labor Impact

Let's talk numbers. In the US alone:

• Forklift operators: 875,000 jobs
• Heavy equipment operators: 1.2 million jobs
• Delivery drivers: 2.8 million jobs (last-mile only)
• Industrial machinery operators: 1.5 million jobs
• Construction equipment operators: 550,000 jobs

That's over 6 million jobs that require spatial reasoning and physical environment navigation. Jobs that have been automation-resistant specifically because AI couldn't handle complex 3D environments.

If General Intuition's approach works - and VCs just bet $134 million that it will - these jobs move from "hard to automate" to "economically automatable."

Why The Massive Seed Round

Seed rounds are usually $1-5 million. Maybe $10 million if you're hot. $134 million for a seed is basically unheard of.

Here's what it signals:

1. The Race Is On

VCs know that whoever cracks spatial AI first will dominate robotics and physical automation for the next decade. General Intuition isn't raising this much money to build a nice product. They're raising enough to win the category before competitors even start.

2. Training Data Advantage

Medal's gaming platform has massive amounts of spatial navigation data that competitors don't have access to. That's a defensible moat in AI, where training data is often the difference between winning and losing.

3. Infrastructure Costs

Training spatial AI models requires enormous compute power. Gaming data is high-dimensional (3D + time). Processing it isn't cheap. This funding gives them the runway to train at scale.

4. Investor Conviction

Khosla Ventures, General Catalyst, and Raine are top-tier firms. They don't do $134M seeds unless they see category-defining potential. This is them signaling to the market: spatial AI is the next frontier.

Who's At Risk

If you operate any kind of vehicle, equipment, or machinery in a dynamic environment, your job just moved higher on the automation risk list:

• Equipment operators: Your job requires spatial reasoning. That's exactly what this tech is being built to replace.
• Drone pilots: Once drones can navigate autonomously in complex environments, the pilot becomes optional.
• Warehouse workers: Physical navigation is the main barrier keeping some warehouse jobs human. That barrier is eroding.
• Delivery drivers: Last-mile delivery needs spatial reasoning for sidewalks, buildings, obstacles. Spatial AI solves that.
• Construction workers: Operating equipment currently requires human judgment about terrain, obstacles, and spatial planning.

The timeline? Probably not immediate. Spatial AI needs to prove itself in controlled environments first, then graduate to more complex scenarios. But we're talking 3-5 years until commercial deployments, not 10-20 years.

What You Should Know

Look, this is early-stage tech. General Intuition could fail. The approach might not work at scale. There are massive regulatory and safety hurdles for autonomous systems in physical industries.

But here's what's undeniable:

Major VCs just bet $134 million that AI can learn to navigate the physical world by playing video games. That's not a small bet. That's "we think this changes everything" money.

If they're right - even partially right - a huge category of jobs that seemed safe from automation because they required physical spatial reasoning suddenly becomes vulnerable.

The good news? You have time to prepare. Spatial AI isn't deployed at scale yet. But the funding is there. The tech direction is set. The applications are clear.

For operators and workers in physical industries:

• Monitor your industry for pilot programs and automation initiatives
• Move toward supervisory roles where you oversee autonomous systems rather than operate manually
• Develop skills in AI system management, maintenance, and troubleshooting
• Unionize if you haven't already - collective bargaining can slow adoption and ensure worker protections

Or assume you're safe because your job requires navigating physical space. That worked great until a gaming company raised $134 million specifically to automate that skill.

Your call.

← Back to News