Chat GPT’s Free Status: What You NEED to Know


Do you remember when Google was just starting out? Despite being small and facing huge expenses, it managed to gain the support of major players because it had a bright future ahead. Similarly, ChatGPT’s owner, OpenAI, has created a product that has the potential to be a major profit center but is costing them 3 million a month or more. So, are you wondering if ChatGPT will remain free? Here’s my prediction.

OpenAI CEO Sam Altman has revealed plans to monetize Chat GPT. This could take the form of advertising, akin to Google’s model, a licensing model, or a hybrid. These approaches would enable Chat GPT to remain free for users.

So what method of monetization will Open AI use with Chat GPT? We have a hint from their charter and actions but let’s take a look.

Table of Contents

A Portion of the OpenAI Charter Source https://openai.com/charter/

Does Chat GPT Monetizing Mean it Won’t Be Free?

Will Chat GPT monetize? Yes according to Sam Altman one of the founders and CEO of OpenAI that owns Chat GPT and the underlying technology.

Twitter Conversation : Sam Altman CEO OpenAI on monetization of Chat GPT

So what does it mean to monetize for Chat GPT?

Here are most of the options for Chat GPT monetization.

  • Licensing: Charging other businesses or organizations a fee to use Chat GPT technology in their own products or services.
  • Advertising: Integrating advertising into the Chat GPT platform and generating revenue through ad clicks or impressions.
  • Subscription model: Charging users a recurring fee to access Chat GPT’s features and functionality.
  • Partnerships and sponsorships: Forming partnerships or securing sponsorships from other businesses as a source of revenue.

Licensing:

Licencing is an interesting option for OpenAI for the technology that underlies Chat GPT.

If for instance, if OpenAI licensed the name “Chat GPT” and the underlying technology they could simply charge their own spin-off company or a licensee a fee. That fits with how they have been doing business so far.

This model would also let them license the technology to other companies that could compete against each other. That has the potential at least to,

  • Allow the technology to benefit the maximum number of people
  • Keep too much power from accruing in the hands of a single company
  • Move the costs of the computing power over to a third party
  • Allow them to focus on what they do best, develop the technology
  • Keep OpenAI at the forefront of Chat Bot development

The licensing model is the most likely monetization method based on how OpenAI has acted in the past and its charter.

Advertising:

Retaining Chat GPT and the underlying technology for themselves and setting up a pay per click advertising model like Google has is probably the most lucrative way of monetization for OpenAI. The problem is that will tend to concentrate a significant amount of power into their hands.

  • Based on the OpenAI charter they do not want to become the next Google
  • The advertising model would quickly lead to conflicts of interest between users, stakeholders, and employees. This is also against the OpenAI charter.

Subscription model:

If OpenAI monetized Chat GPT using a subscription model it would certainly be lucrative, but with a lot of downsides.

  • Subscriptions would force people to pay for Chat GPT, the percentage of users willing to pay is questionable and will be dependent on the price.
  • Many people in third-world countries would be priced out of the service unless special pricing arrangements were made for them.
    • Many people in poorer countries do not have banking available.
  • Limiting AI to only the “rich” goes against the OpenAI charter.

Partnerships and sponsorships:

It is easy to imagine Google partnering with OpenAI to make Chat GPT exclusively available to their customers. The problem with this

  • It would go against the OpenAI charter by concentrating even more power in Google’s hands.

It’s possible that Google might license ChatGPT but it is unlikely that it would be an exclusive agreement.

  • Sponsorships are possible, but sponsorships usually end up being semi-exclusive advertising agreements.
  • This could work well for both parties but, it does risk concentrating too much power in OpenAI’s hands.
  • Google has its own AI technology as well called Lambda, it could potentially be ramped up to be a real competitor to Chat GPT.

Overall licensing seems like the most likely way that Chat GPT will be monetized.

How much does ChatGPT cost to run?

Tom Goldstein has done a “back of the envelope” calculation. This estimate is the best number that I have found so far.

Tom Goldstein Twitter Thread Calculating Chat GPT Cost to Run (Estimate)

So according to the estimate Chat GPT costs about $3 Million Dollars a month to run.

This fails to take into account that Chat GPT has gone viral and there are probably a lot more than 1 million users currently.

According to Sam Altman in a Twitter chat with Elon Musk (co-founder of Chat GPT). The cost of each chat is “average is probably single-digits cents per chat; trying to figure out more precisely and also how we can optimize it”

Chat with Elon Musk Co-founder of OpenAI Twitter Chat

Another Estimate Based on What OpenAI Actually Charges

We can also look to see what OpenAI is currently charging to use their GPT-3 model to generate content.

Doing the math for the cost for 1000 words is about,

  • 2.6 cents for the most expensive model (Davinci)
  • .053 cents for the cheapest (Ada)

So if we estimate that the cost is closer to Davinci than Ada and that they are making a 50% margin.

  • We can estimate the cost for Chat GPT is about 1.5 cents per 1000 words.

Looking at my own chats, (I’ve been doing a lot of research)

  • My average chat length is about 2,600 words
  • The longest chat was 5,862 words and the shortest was 818 words.
  • That is a pretty large variance.

So I can estimate that my own chats are costing OpenAI about

2,600 words * 1.5 cents/1000 words = 3.9 cents per chat (estimated average cost per chat)

I am having about 3 chats a day with Chat GPT so I am costing OpenAI about,

3.9 cents * 3 chats/day = 11.7 cents /day

If I am an “average user”, then the cost for 1 Million users per day is,

11.7 cents / day * 1 million users = $117,000 dollars / day

That is remarkably close to Tom Goldstein’s estimate of $100,000 a day.

What we really don’t know is how many users there are currently using Chat GPT. We can estimate that though assuming exponential (viral) growth. Running that estimate though gives me literally everyone in the world. (https://www.stepbystepsolutioncreator.com/st/ex using 1 for day 1 and 1,000,000 for day 5, currently, we are at day 28).

If that is right they could be spending 100 million dollars a day for 1 billion users. I think that is not very likely or they have remarkably deep pockets. It does point out though that their costs quickly scale to an enormous amount of money.

Are There Other Free Chat Bots Like Chat GPT?

The short answer to this is not yet, although there are some chatbots out there that can answer a few questions they are dumb compared to Chat GPTF. Let’s experiment with a couple of them.

Blender Bothttps://blenderbot.ai/chat

Blender AI Response to Source Query

Chat GPT makes up sources for this query, (one of its downsides), but at least it answers the question.

Asking Blender Bot to give me some pseudocode. ChatGPT does this with no issues.

Overall a weak showing compared to Chat GPT.

You Chathttps://you.com/

You Chat: None of These Links Work

You chat seems to be trying to answer the query. The sources are all places that could have information about how to tell a horse’s age by its teeth, but… None of these links work.

So far I have not found anything close to Chat GPT. Please leave me a note in the comments if you have something I should test.

Chris

Chris Chenault trained as a physicist at NMSU and did his doctoral work in biophysics at Emory. After studying medicine but deciding not to pursue an MD at Emory medical school Chris started a successful online business. In the past 10 years Chris's interests and studies have been focused on AI as applied to search engines, and LLM models. He has spent more than a thousand hours studying ChatGPT, GPT 3.5, and GPT4. He is currently working on a research paper on AI hallucinations and reducing their effects in large language models.

Recent Posts