As of September 2025, the AI economy is at a pivotal juncture, marked by rapid advancements and mounting challenges.
The cost of generating one million tokens has dropped approximately 100-fold, yet surging demand outpaces this decline.
Two years ago, SaaS companies handled 1-2k tokens; now, a standard deep research query requires 1 million tokens, with projections of 10-100 million per query by next year.
Notion, once enjoying a 90% margin typical of cloud SaaS, is seeing this erode as CEO Ivan Zhao reveals AI provider payments now account for about 10% of its revenue.
Similarly, Cursor, Replit, and others are raising prices despite 80-90% margins from large clients, prompting a shift to cost-effective models like GPT-5 Nano ($0.10 per million tokens) compared to GPT-5’s $3.44.
Industry giants — Google, OpenAI, and Anthropic — dominate infrastructure, making it increasingly difficult for startups to compete. Despite promises of cheaper AI, the real economic burden is rising, straining startup business models. This benefits clients with improved efficiency but risks driving market centralization.
Also read:
- Copilot Arrives on Samsung TVs and Monitors
- Andreessen Horowitz Releases Fifth Top 100 AI Apps Ranking
- China Unveils "AI Plus" Strategy for Total Economic Integration
Thank you!
Join us on social media!
See you!
Author: Slava Vasipenok
Founder and CEO of QUASA (quasa.io) - Daily insights on Web3, AI, Crypto, and Freelance. Stay updated on finance, technology trends, and creator tools - with sources and real value.
Innovative entrepreneur with over 20 years of experience in IT, fintech, and blockchain. Specializes in decentralized solutions for freelancing, helping to overcome the barriers of traditional finance, especially in developing regions.
This is not financial or investment advice. Always do your own research (DYOR).
