EMEmergent MotionContact us

AI for dynamic worlds

Intelligencein Motion.

We build AI models that understand movement, decisions and outcomes in dynamic environments.

From sports to real-world systems.

Our mission

To advance the frontier of AIby teaching machines to perceive,reason and act in complex,dynamic worlds.

We believe the best way to build general intelligence is by learning from environments where decisions matter.

Why sports?

Sport is
intelligence
in motion.

Sports create some of the richest environments for studying intelligent behaviour.

Athletes continuously perceive their surroundings, anticipate the actions of others, evaluate possibilities and make decisions under extreme time pressure.

movement × space × intention × action × outcome

We use these environments to train models that learn how dynamic worlds evolve.

Sports are the beginning. Not the boundary.

Our first environments

Intelligence behaves differently in every world.

Domain

EMERGENT
FOOTBALL

Understanding the beautiful game through movement, space, decisions and context.

Floodlit football pitch with professional players in dark kits

State: observing

Player 8

PressureHigh
Options4
Best actionForward pass
Decision value+18.4%

Domain

EMERGENT
RACING

Modeling high-speed decisions where fractions of a second change everything.

Black Formula-style race car cornering on a night circuit

Strategy

Pit now18%
Lap +127%
Lap +255%
RecommendationExtend 2 laps

Domain

EMERGENT
CYCLING

Optimizing performance across terrain, tactics and team dynamics.

Elite road cyclists climbing a mountain road at dusk
GRADIENT 7.8%
Attack18%
Hold61%
Drop back21%

Rider 21

Power418 W
Gradient7.8%
Draft benefit12%
Energy reserve74%

Domain

EMERGENT
BASEBALL

Modeling pitch, swing and fielding decisions across every plate appearance.

Baseball pitcher mid-delivery in a floodlit night stadium

Pitch 42

Velocity96.4 mph
Spin2,410 rpm
Zone edgeLow away
Whiff prob.31%

Domain

EMERGENT
BASKETBALL

Reading spacing, off-ball movement and shot selection in continuous possession.

Basketball player rising to the rim in a darkened arena

State: observing

Player 23

SpacingOptimal
Shot quality0.62
Best actionDrive right
Decision value+11.2%

Domain

EMERGENT
AMERICAN FOOTBALL

Understanding play design, coverage and outcome value snap by snap.

American football teams set at the line of scrimmage at night

State: observing

Play 3rd & 6

CoverageCover 3
Pressure prob.38%
Best readSlot seam
Conversion+14.7%

Live simulation

A decision, modelled in real time.

The simulation pauses at a decision point, expands the available actions, scores each continuation, then executes the model-selected option.

Option A

64%

Option B

21%

Option C

15%

State: observing

Our platform

From raw data to real-world intelligence.

  1. 01

    CAPTURE

    Video, tracking and telemetry data from the real world.

    Video pixels

  2. 02

    RECONSTRUCT

    Rebuild the complete state of the environment in space and time.

    Point cloud

  3. 03

    UNDERSTAND

    Identify actions, intentions and interactions.

    Network graph

  4. 04

    SIMULATE

    Simulate thousands of possible futures and counterfactuals.

    Branching trajectories

  5. 05

    DECIDE

    Evaluate possibilities and identify optimal decisions.

    Selected branch

  6. 06

    IMPACT

    Turn intelligence into better performance, strategy and understanding.

    Target pulse

One state.
Thousands
of futures.

Our models don't only analyze what happened. They learn what could happen next.

CURRENT STATE
Observe
Simulate
Predict
Decide

Counterfactual intelligence

What if?

Change the action taken at 87:32 and the model re-simulates every future that follows from it.

Match clock87:32
STATE(t) + ACTION(t) → STATE(t+1)

Possession lost.

Select an alternative action

Possession retained
31%
Chance creation
12%
Field progression
24%
Risk
78%

World models

Learning how dynamic worlds evolve.

Emergent Motion transforms real-world movement into structured representations of state, action and outcome.

STATE(t) + ACTION(t) → STATE(t+1)

StatePossible actionsPossible futuresOptimal decision

Built for the world's
most ambitious teams.

We work with teams, athletes, researchers and technology partners pushing the limits of human and machine performance.

Partner with us

The world never stops moving.

Neither does intelligence.

Emergent Motion

Building AI that understands what happens next.