AI Competitive Dynamics

The competitive dynamics of AI are unlike most previous technology waves. Understanding the structural forces shaping AI competition is important for strategic planning.

The foundation model layer

A small number of organisations (OpenAI, Anthropic, Google, Meta, Mistral) are building and training frontier models. This requires compute at the scale of billions of dollars and talent that is genuinely scarce globally. The competitive dynamics at this layer favour large incumbents. Most organisations should not try to compete here.

The application and customisation layer

The vast majority of AI value for most organisations lies in the application layer — building products and workflows on top of foundation models. This layer has lower barriers and higher competition. Sustainable advantage comes from: proprietary data, domain expertise, distribution advantages, and switching costs built into workflows.

Disruption risk

AI may lower the barriers to entry in industries where incumbent advantage has been built on information asymmetry, scale advantages in knowledge work, or high switching costs in software. Organisations should honestly assess where AI might enable new entrants or smaller competitors to credibly compete with them.

The NZ competitive context

New Zealand organisations face the structural challenge of being price-takers in foundation models (dominated by US and increasingly Chinese companies) while being well-positioned to use AI to overcome the productivity and scale limitations that come from operating in a small market. The strategic opportunity is AI-enabled globalisation of NZ services and products.

The competitive question for most NZ organisations is not "how do we build better AI than Google?" It is "how do we use AI to compete internationally with the unique knowledge, data, and capabilities we already have?"