Running artificial intelligence (AI) models ain’t cheap. OpenAI 1st released ChatGPT to the public on a trial basis at no cost, but it costs them plenty; reportedly a few cents per question. If you consider that millions of people are interacting with ChatGPT on a daily basis then you can figure out how deep OpenAI’s pockets need to be.
Monetizing is a priority for OpenAI. It’s under pressure from Microsoft to turn a profit after the tech giant invested over $10 billion in its product. Judging by Tweets on Twitter, one of the key areas of focus will be to offer a dedicated computing platform to serious developers at a profit. This platform will be called “Foundry”.
Foundry will facilitate large workloads & offer clients complete control over the configuration of the model. Clients will be able to use OpenAI’s tools & dashboard to develop models. New releases of GPT, such as the much-anticipated version 4 will, automatically, be on offer to clients using Foundry.
All Foundry’s clients will be able to rely on guaranteed uptime of 99.5% for their instances & dedicated support from OpenAI’s engineering team. Clients will also be given absolute control of the snapshot instance they want to run, giving them leeway to develop robust models on top of OpenAI.
Foundry’s hefty fees will not pull OpenAI’s balance sheet into the black: predictions estimate that the company is set to earn $200 000 this year & $1 billion next year. These predictions are based on anticipated earnings from ChatGPT Plus, which costs $20 per month.
Image credit: Twitter
