General Intuition's $6B World Model: How Simulation-Based AI Is Reshaping Enterprise Planning

General Intuition's $6B World Model: How Simulation-Based AI Is Reshaping Enterprise Planning

General Intuition, a New York startup spun out of gameplay-clip platform technology, nearly tripled its valuation from $2.3 billion to $6 billion in just 8 weeks on a $320 million round led by Valor Equity Partners, Point72 Ventures, and Seven Seven Six. The company builds world models — simulation engines that predict the consequences of actions before they are taken — and their enterprise customers report 71% fewer planning errors after deployment. This is not incremental improvement; it is a fundamental shift in how enterprises make decisions.

The world model approach inverts the traditional enterprise planning workflow. Instead of proposing an action, executing it, measuring results, and iterating, enterprises now propose actions to the world model, receive predicted outcomes with confidence scores, and only execute when the simulation predicts acceptable results. The analogy is the difference between crash-testing a car by actually crashing it versus running the crash in a physics simulator first.

What a World Model Actually Does

A world model is a learned representation of how systems evolve over time given inputs. Unlike a traditional ML model that maps input to output, a world model maps input to a predicted trajectory of states.

Traditional ML:  Input → Output (one prediction)
World Model:     Input → State₁ → State₂ → State₃ → ... → Stateₙ
                 (predicted trajectory of outcomes)

For enterprise planning, this means:

Scenario Simulation. "If we raise prices by 12% and reduce marketing spend by 20%, what happens to revenue, churn, and market share over the next 6 months?" The world model simulates the trajectory rather than making a single-point prediction.

Causal Inference. "Did the price increase cause the churn spike, or was it the marketing reduction?" The world model can run counterfactual simulations — "What if we had raised prices but kept marketing spend?" — to isolate causal effects.

Risk Quantification. "What is the probability that this pricing strategy results in >5% churn?" Monte Carlo simulation across the world model's predicted trajectories provides statistically grounded risk estimates.

The $6B Valuation Math

The 24-33x revenue multiple is high but not unprecedented for infrastructure AI companies. Datadog traded at 25x revenue at its peak, and Palantir at 30x. General Intuition's argument is that world models become the operating system layer for enterprise decision-making — every decision passes through simulation, creating an API call per decision rather than an API call per query.

How Enterprises Are Using World Models

Supply Chain Optimization. A Fortune 100 retailer uses General Intuition to simulate the impact of supplier changes, tariff adjustments, and demand shifts on their supply chain before committing to contracts. The simulation runs 500 scenarios in 10 minutes, compared to 2 weeks of manual analysis.

Pricing Strategy. A SaaS company simulates the revenue, churn, and competitive impact of 20 pricing configurations before launching an A/B test. The world model identifies 3 configurations with >80% probability of positive revenue impact, reducing the A/B test surface from 20 to 3 variants.

Hiring and Team Building. A tech company simulates the productivity impact of different team compositions before making hiring decisions, predicting which candidate combinations will produce the highest team velocity.

The Technical Architecture

General Intuition's world model combines three architectures:

Transformer-based state prediction. A large transformer model learns state transitions from historical enterprise data, predicting how system state evolves given interventions.

Graph neural networks for causality. A GNN layer learns the causal structure of enterprise systems — which variables affect which, with what delay and magnitude.

Monte Carlo sampling. Multiple simulations with stochastic perturbation provide confidence intervals and risk estimates rather than single-point predictions.

The Competitive Landscape

General Intuition's moat is the learned world model itself: every enterprise interaction improves the simulation, creating a flywheel that is difficult to replicate. A company that has simulated 10 million pricing scenarios has a fundamentally better pricing world model than a competitor starting from zero.

Key Takeaways

• General Intuition's world model technology reduces enterprise planning errors by 71% by simulating outcomes before execution, inverting the traditional trial-and-error workflow
• The $6B valuation reflects a 24-33x revenue multiple on an estimated $180-250M ARR, justified by the API-call-per-decision business model that scales with enterprise decision volume
• The learned world model creates a competitive flywheel: every enterprise interaction improves simulation accuracy, making it increasingly difficult for competitors to match prediction quality

By Deepak Bagada, CEO at SaaSNext & Principal AI Architect.

Last tested: August 2026 with Python 3.12, Node v22, and latest framework releases.