I’m an enthusiastic user of Cursor and incredibly excited about the immense potential AI brings to programming and many other fields. However, there’s a critical issue I believe we need to discuss: the current pricing structure of AI services and its fairness for users in countries with vastly different income levels.
We all know there’s a significant income disparity between regions like Europe or North America and Southeast Asian countries (like my home country of Vietnam, or others like Indonesia and the Philippines). For instance, a $20/month subscription for a powerful AI tool might be a negligible expense for a developer in Europe, but it can represent a substantial portion of a developer’s monthly income in Southeast Asia.
Most powerful AI services, including those deeply integrated into Cursor (like GPT-4), often operate on a global, one-size-fits-all pricing model. This means that, in terms of real income proportion, users in developing countries are effectively paying a disproportionately higher price to access and utilize the same technology.
This raises several important questions and concerns:
Accessibility Barrier: Is the current pricing inadvertently creating a barrier, making it harder for talent and businesses in lower-income countries to access and leverage AI’s power?
Fairness: Is a flat global price truly equitable when purchasing power and economic conditions vary so dramatically across regions?
Stifling Innovation: If AI becomes prohibitively expensive for a large segment of the global population, could this slow down innovation and AI adoption in developing nations, potentially widening the digital divide?
I believe that considering models like regional pricing or pricing based on purchasing power parity (PPP) would be a more equitable and humane approach. Many major software and service companies (e.g., Netflix, Spotify, Steam) have successfully implemented such policies, making their products accessible to a much wider global audience.
I’d love to hear the thoughts of the Cursor community on this, especially from users in different regions, and hopefully from the Cursor development team as well:
What are your thoughts on this issue?
Could Cursor, or AI providers in general, consider more flexible pricing policies for different markets?
Are there other viable solutions to make AI more accessible and equitable for everyone, regardless of their geographical location or economic standing?
Thanks for reading, and I look forward to your valuable insights!
Each nation is free to develop their own national or cultural locked AI models.
The big model companies are as fair as possible, you can pay a bite of bread for a prompt.
Local models exist, they only cost the existing hardware available.
These tools and this technology can actually solve issues, with huge monetary impact. If you cannot use the tools economically, don’t. Spend your entertainment money instead, when you consider it to be entertaining.
Find educational paths that are able to give you access to LLMs, like schools, grants, bourses, wealthy people around you, etc.
20$/month is TOO MUCH for 90% of the world.
20$/month is INSIGNIFICANT for 10% of the world.
Many things are like this in the world.
You can use Cursor today, with the best models, non-stop, and spend over 1000000$/year at current prices. You can also use free tools, or cheap tools, to learn how to be more effective in using them cheaply, and still solve some real world problems that can bring in money.
When the Internet was first starting to become popular in my country, I spent 50% of a monthly salary to go online, at night, when it was cheaper. The result is that I had access to knowledge and experiences that were not possible otherwise and I was professionally prepared to participate in a global market after the dot com bubble.
We are now seeing the AI bubble forming with huge promises and all eyes looking. You being on this forum puts you above 90% of the people around you in the next years, in both being able to take opportunities but also avoid risks. This is bigger than the dot com.
I do truly appreciate the desire to make Cursor more accessible in other countries. Unfortunately, we have direct costs to the LLM providers that we have to pay per request, making it hard to offer such discounts.
Companies like Netflix and Spotify don’t have this issue, as they own and run all of their infrastructure, so the cost to stream a song is controlled by them, so they choose the price.
We have always made decisions to keep the Pro plan at $20/m to make it as accessible as possible, but any lower makes the pricing model not viable.
We have recently upgraded our free tier, so you get 500 free model requests, and 50 premium requests a month, plus 2000 Tab completions, which do actually renew each month instead of being a one-off allowance like in the past.
We hope to be able to bring Cursor to more people in the future
Why don’t you guys try to partner with Claude, like how Microsoft Azure deploys the ChatGPT model on their servers and doesn’t have to pay for the API?
