AI BriefingGenerative AI Economics: Changes After 2 Years
Key point
Even though the generative AI market has grown 5x, most of the profit is still monopolized by the semiconductor layer.
Details
Over the past 2 years, the generative AI ecosystem has grown roughly 5x, reaching $435 billion in annual revenue, but the economic structure has not changed much. The value chain remains inverted, with the Semiconductor (Semi) layer still accounting for about 70% of total revenue.
The breakdown by layer is as follows:
- Semiconductor (Semi): About $300 billion in size, with NVIDIA monopolizing roughly 80% of this market.
- Infra: About $75 billion in size, relatively evenly distributed among major cloud companies such as Azure, AWS, and GCP.
- Apps: About $60 billion in size, with OpenAI and Anthropic holding roughly 75% of the market.
The gap is also stark in terms of profitability. The semiconductor layer takes about 79% of the entire AI ecosystem's Gross Profit. This is the exact opposite of the traditional cloud model, in which applications capture 70% of the profit.
In conclusion, semiconductors show monopoly market characteristics, apps show a duopoly structure, and infra remains a competitive layer. The most profitable strategy in the AI industry is still selling 'picks and shovels' (infra and semiconductors).
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Anthropic Pacing to $100B ARR; Rumors of OpenAI Solving Navier-Stokes Spark ControversyOpenAI and Anthropic's RSI Strategy and the Uncertainty of AI Profitability