AI GAME THEORY
SERIES 00 / VERSION 0.01
INTERACTIVE ANALYSIS

AI GAME
THEORY

HOW TO USEBUILD A SCENARIO
  1. 01

    Choose what is scarce. Set the capability or chokepoint that governs bargaining power.

  2. 02

    Select the players. Choose one for a strategic dossier or two for a bilateral game.

  3. 03

    Change the conditions. Compare how incentives, leverage, and likely outcomes move.

Repeated bargaining loopControl of a scarce asset and relative outside options determine the focal player's bargaining power. Bargaining power determines who captures current surplus and retains proprietary learning. Both feed back into the next round.CONTROL OFSCARCEASSETRELATIVEOUTSIDEOPTIONSCURRENTSURPLUSCAPTUREDPROPRIETARYLEARNINGRETAINEDFOCAL PLAYERBARGAININGPOWER
FIG 1.0 — REPEATED BARGAINING LOOPCONTROL + RELATIVE OUTSIDE OPTIONS DETERMINE POWER
POWER DETERMINES WHO CAPTURES SURPLUS + RETAINS LEARNING
CENTRAL THESIS

The AI economy is a repeated bargaining game over two prizes: current margin and future learning. Each player wants to own one scarce bottleneck, make its dependencies interchangeable, and retain the learning generated by real usage.

RELATIONSHIP ENGINESELECTION LIMIT: 02

PLAYER
FIELD GUIDE

Inspect what each player wants, what it fears, when leverage rises or falls, and the major moves it makes to remain strategically dominant.

SELECT ONE MORE PLAYER TO COMPARE
ChipsPLAYER 2
CONTROL POINTWhich control point governs the game?
CROSS-CUTTING PLAYER / NOT A LAYER
DATA + LEARNING are the state variable: every transaction changes the next round
PLAYER L-01 / STRATEGIC DOSSIER

Chips & systems

Commoditize intelligence; preserve computation.

01

Strategic position

The chip company’s ideal world contains thousands of capable models, millions of applications, and no model provider powerful enough to dictate infrastructure economics.

Models are complements to compute. More models create more training; cheaper models create more deployment. The objective is not openness in the abstract, but abundant demand for a scarce computing platform.

KEEPS SCARCEEfficient compute and the software ecosystem around it
WANTS COMMODITIZEDModels and applications
MARKET REGIME ICapability scarcityA clear model-quality gap

Chips does not control the resulting bottleneck: Frontier capability. When the best models are meaningfully better than alternatives, frontier labs capture the surplus and downstream players tolerate dependency.

02

What chips want

Fragmented model demand

Many independent model builders broaden the market for compute and prevent one customer from dictating infrastructure economics.

Elastic inference growth

Usage must expand faster than efficiency reduces compute per task, or cheaper intelligence will compress hardware demand.

Technical instability

Rapid architectural change preserves the value of flexible, programmable accelerators over workload-specific custom silicon.

Platform dependence

Customers should build around the chip company’s software, networking, and systems so switching hardware remains costly.

Distributed buyer power

No frontier lab or cloud should become large enough to act as a dominant buyer and negotiate away the supplier’s margin.

03

What chips fear

Custom silicon

At sufficient scale, a hyperscaler can optimize an accelerator around recurring workload shapes and reduce its exposure to general-purpose GPU economics. Google TPUs and AWS Trainium show this is already a deployed strategy; the constraints are workload fit, software maturity, and utilization—not whether custom silicon exists.

Algorithmic efficiency

More value per unit of computation does not automatically increase chip demand. Consumption must expand quickly enough to offset falling compute per task.

Buyer concentration

A few large labs and clouds create a monopsony problem. They can negotiate aggressively, fund substitutes, and determine the shape of advanced demand.

04

How leverage changes

LEVERAGE IS HIGHEST WHEN
  • Advanced capacity is scarce
  • Model architecture changes quickly
  • Software switching costs remain high
  • Demand is distributed across many buyers
LEVERAGE FALLS WHEN
  • Supply catches up
  • A few clouds dominate procurement
  • Inference becomes standardized
  • Custom accelerators become competitive
  • Model efficiency grows faster than demand
05

The major move

Move beyond the chip into networking, systems, software, models, and inference—defense against the possibility that the accelerator itself becomes the replaceable layer.

DEPENDENCY / HOLD-UP RISK

A small number of clouds and labs concentrate procurement and can fund substitutes

OUTSIDE OPTION

Move into networking, systems, software, models, and inference

06

Proprietary learning

Chip companies see which model operations slow down, which parts of their systems are fully used, and which workloads are growing. They use that evidence to decide what the next chips, networks, and software should improve.

07

Who plays this role

PUBLIC PHILOSOPHY

Open models and abundant AI adoption expand the market for computation.

Select any second player—from either index—to analyze their relationship.